Targeting collateral deletion of housekeeping genes caused by the loss of tumor suppressor genes is a potential strategy to identify context-specific, molecular-targeted therapies in cancer. In mammals, phosphatidylserine (PS) synthesis depends on two redundant PS synthetases, PTDSS1 and PTDSS2, and PTDSS2 is located at a tumor-suppressive locus, 11p15.5. Here, we sought to determine whether PTDSS2 loss would confer vulnerability to disruption of PTDSS1 function. PTDSS2 was lost in a wide range of cancer types, and PTDSS1 depletion specifically suppressed growth in PTDSS2-deficient cancer cell lines. Potent and selective PTDSS1 inhibitors were developed whose activity phenocopied the effect of PTDSS1 depletion, and in vivo treatment of PTDSS2-deleted tumors with these inhibitors led to tumor regression. Whole-transcriptome analysis revealed that inhibition of PTDSS1 in PTDSS2-depleted cells activated the endoplasmic reticulum (ER) stress response downstream of PS depletion. ER stress-mediated cell death induced by PTDSS1 inhibitors activated tumor immunity through the secretion of HMGB1 protein followed by activation of bone marrow-derived dendritic cells. PTDSS2 loss showed intratumoral heterogeneity in clinical samples, raising concerns about resistance to PTDSS1 inhibition. However, the PTDSS1 inhibitor effectively suppressed the growth of tumor containing both PTDSS2 wild-type and knockout cells in immunocompetent mice, showing potency for overcoming tumor heterogeneity by modulating the tumor immune microenvironment. Thus, these newly developed PTDSS1 inhibitors provide a therapeutic option for treating cancer with PTDSS2 loss, harnessing the synthetic lethality of PTDSS1/2.

Significance:

This study identifies a specific dependency on PTDSS1 for phosphatidylserine synthesis following PTDSS2 deletion and introduces novel PTDSS1 inhibitors as a therapeutic option to induce collateral lethality in cancer with PTDSS2 loss.

Targeting undruggable loss-of-function mutations via synthetic lethal strategies has received increasing attention as a potential strategy to identify context-specific molecular-targeted therapies in cancer. A celebrated proof-of-concept example for the targeting of loss-of-function mutations is the clinical effectiveness of PARP inhibitors in cancers with BRCA gene mutations and alterations (1, 2). Targeting deletion of functional genes as collateral damage of tumor suppressors has been a successful approach to reveal novel synthetic lethal pairs, exemplified by the first discovery of ENO2 as a synthetic lethal target of ENO1-deleted tumor (3, 4), and the subsequent discovery of PRMT5 as a synthetic lethal target of MTAP1 deletion, which is codeleted with CDKN2A (5, 6). In addition, screening of a natural compound library revealed thymoquinone and phorbol ester as synthetic lethal inducers in prostate cancers with RB1-SUCLA2 deletion (7).

Phosphatidylserine (PS) is an essential phospholipid that constitutes the cellular membrane. PS synthesis in mammalian cells depends on two ubiquitously expressed enzymes, PTDSS1 and PTDSS2 (8–10). PTDSS1 produces PS from phosphatidylcholine (PC), whereas PTDSS2 does from phosphatidylethanolamine (PE) through a base exchange reaction. Single knockout of either PTDSS1 or PTDSS2 did not affect mouse viability, but PTDSS1 and PTDSS2 double knockout resulted in embryonic lethality, suggesting PS is essential for cell viability and PTDSS1 and PTDSS2 function redundantly to produce PS (8, 9). PTDSS2 is located at a tumor-suppressive locus, 11p15.5, which is frequently lost in a variety of cancer types including lung and breast cancer, and glioma (11–16). These observations inspired us to explore the potential of PTDSS1 as a synthetic lethal target for PTDSS2-deleted tumor cells.

Here, we show that PTDSS2 is lost in human cancer cells by clinical sample analysis and confirm the conserved synthetic lethal relationship between PTDSS1 and PTDSS2 in human cancer cells. We have developed potent and selective PTDSS1 inhibitors, which have a clear safety profile and show selective killing of PTDSS2-deleted cancer cells in vitro and in vivo. Transcriptomic analysis revealed that the efficacy stems from selective activation of ER stress by PTDSS1/2 double inhibition. We further revealed that the ER stress induced by this synthetic lethality could activate antitumor immunity through the secretion of damage-associated molecular pattern molecules (DAMP) and the subsequent activation of dendritic cells. Our discovery provides a therapeutic opportunity for patients with PTDSS2-deleted tumors by utilizing the synthetic lethality between PTDSS1 and PTDSS2.

Data collection and ethical considerations

Clinical and histopathologic data were collected retrospectively in a nonstratified manner. Data included patients’ age, primary cancer site, TNM stage, disease-specific survival, histologic subtype, presence of vascular invasion, and tumor diameter and grade. Approval for the use of samples and the patient clinical background data required for the analysis and tissue collection was obtained from the local ethics committee [Ethikkommission Norwest- und Zentralschweiz (EKNZ) Permission 2017-00302]. All experimental procedures using animals were approved by the in-house guidelines of the Institutional Animal Care and Use Committee of Daiichi Sankyo Co., Ltd.

Tissue microarray

A pan-cancer tissue microarray (TMA) dataset was established from specimens at the Biobank at the Institute of Pathology, University Hospital Basel, Switzerland. Thirty-one TMA blocks of nonconsecutive primary cancer specimens and ≈30% paired nonmalignant adjacent tissue specimens were constructed using TMA-Grand Master (3DHisteck; Sysmex) as formalin-fixed, paraffin-embedded tissue blocks, prepared following standard protocols (17). The percentage of tumor cells within each core sample was >50% for all tumor specimens. Of the 31 TMAs comprising 2,988 cores from 2,492 patients, 17 cancer types, each with more than 30 patients, were selected to evaluate PTDSS2 status.

Cell lines and reagents

HCT116 (ATCC, catalog no. CCL-247, RRID:CVCL_0291), A375 (ATCC, catalog no. CRL-1619, RRID:CVCL_0132), ARPE-19 (ATCC, catalog no. CRL-2302, RRID:CVCL_0145), and CT26.WT (ATCC, catalog no. CRL-2638, RRID:CVCL_7256) cells were purchased from ATCC. PTDSS2-knockout HCT116 clones (#3 and #16) were established with pCas Guide vector with target sequence (5′-CAAGCCACCGGGCCGGGCGA-3′) and pUC vector containing left (exon1, 1–176) and right (exon1, 311–358) homologous arms and GFP-loxP-PGK-puro-loxP functional cassette. The vectors were transfected into HCT116 cells with Lipofectamine LTX Reagent (Thermo Fisher Scientific). Transfected cells were passaged (1:10) until P7, followed by selection with 2 μg/mL puromycin for more than 10 days, after which, single-cell clones were obtained. PTDSS2 heterozygous knockout A375 clone (#27) and homozygous knockout clone (#54) were established with GeneArt CRISPR Nuclease CD4 vector (target sequence: 5′-CGGGCCGAAACGCCATGCGGAGG-3′). The vector was transfected into A375 with Lipofectamine 3000. CD4-positive A375 cells were enriched with Dynabeads CD4 and then single-cell clones were obtained. To confirm the knockout, genomic DNA of these cell lines was purified with DNeasy Blood and Tissue Kit (69506; Qiagen) and PCR was conducted with Multiplex PCR Assay Kit Ver. 2 (RR062A; Takara Bio). Primers used were 5′-ACAATGCACCGCACACCCTTT-3′ and 5′-TGCTGCCACACCCTCAAACC-3′ for the analysis of HCT116-derived cell lines; and 5′-GTGACCCGGTGTGCGTG-3′ and 5′-CTGCCCTCACCAGAAGAAGG-3′ for the analysis of A375-derived cell lines. The PCR temperature cycling conditions were as follows: for HCT116, initial denaturation at 94°C for 1 minute; 30 cycles of denaturation at 94°C for 30 seconds, annealing at 60°C for 1 minute, and elongation at 72°C for 1 minute; and final extension at 72°C for 10 minutes; and for A375, initial denaturation at 94°C for 30 seconds; and 35 cycles of denaturation at 94°C for 30 seconds, and annealing and elongation at 65°C for 30 seconds.

CT26.WT Pss2-KO#1 clone was established with PSS2 CRISPR/Cas9 KO plasmid. Briefly, CT26.WT cells were transfected with the pool of three plasmids with independent guide RNAs (gRNAs; guide A: 5′-TTGTAGCTTACTGGCGGTTT-3′, guide B: 5′-AGTCACGTAGCCCAGCGCAC-3′, guide C: 5′-TTGCCCCCGTAGTCCCTCTC-3′) with Lipofectamine 2000 reagent. The following day, GFP-positive cells were sorted and subjected to single-cell cloning. Pss2 gene knockout was confirmed by RT-PCR. RNA from CT26.WT and CT26.WT Pss2-KO#1 clone was isolated using RNeasy Mini Kit (74106; Qiagen). cDNA was generated using PrimeScript II 1st Strand cDNA Synthesis Kit (6210A; Takara Bio). RT-PCR was performed with Q5 Hot Start High-Fidelity 2× Master Mix (M0494S; New England Biolabs), following the manufacturer's protocol. Primers used were 5′-CTGAGTCTCCGCTGCTCAAG-3′ and 5′-AGCCAATGAAGTGTGCAGGA-3′. PCR products were separated with 2% agarose gel, stained with ethidium bromide, and visualized with E-box gel imager (Vilber Bio Imaging) under ultraviolet light.

IHC and evaluation of PTDSS2 loss prevalence

The staining specificity was first confirmed with PTDSS2-knockout cell lines. HCT116 and A375 parental cells and PTDSS2-KO clones were fixed with 20% neutral-buffered formalin, embedded in agarose gel, and then embedded in paraffin. IHC for paraffin sections of these cell blocks was performed with Ventana BenchMark Ultra (Roche Diagnostics). Anti-PTDSS2 antibody [CEC clone 401-P-17, Lot No. 18E07-IH (NH), raised at Daiichi Sankyo Co., Ltd.] was diluted at 0.5 μg/mL in Ventana Antibody Diluent [251-018 (#95028); Roche Diagnostics]. The staining was visualized with Ventana OptiView DAB Universal Kit [760-700 (#95028); Roche Diagnostics]. The stained slides were digitized with a digital slide scanner, NanoZoomer-XR (Hamamatsu Photonics). IHC for TMA was performed with the same protocol as for the cell blocks.

siRNA transfection and growth assay

The cells were seeded at 1 × 105 cells/well onto a 6-well plate and transfected with 10 nmol/L ON-TARGET Plus SMARTpool siRNA for PTDSS1 (L-008568-00; Dharmacon) or nonsilencing siRNA (D-001810-10-50; Dharmacon) using Lipofectamine RNAiMAX (13778-150; Invitrogen). The transfected cells were seeded at 2 × 104 cells/well onto a 96-well plate. Cell growth was monitored on days 0, 3, 4, 6, and 8 by adding 100 μL/well ATPlite 1step (6016739; PerkinElmer) followed by luminescence detection with EnVision (PerkinElmer).

RT-qPCR

RNA was isolated using RNeasy Mini Kit (74106; Qiagen) or Illustra RNAspin Mini RNA Isolation Kit (25-0500-72; GE Healthcare Japan). DNA was removed by treatment with DNase from the Kit (79254; Qiagen) or with RNase Free DNase I in Illustra RNAspin Mini RNA Isolation Kit. cDNA was generated using High Capacity cDNA Reverse Transcription Kit (4368813; Applied Biosystems). qRT-PCR was performed with TaqMan Fast Advanced Master Mix (4444557; Applied Biosystems) or PrimeTime Gene Expression 2× Master Mix (1055772; Integrated DNA Technologies) following the manufacturer's protocol. TaqMan probes used were as follows: PTDSS1 [Hs00207371_m1 (FAM)], PTDSS2, [Hs00229232_m1 (FAM)], DDIT3 [Hs00358796_g1 (FAM)], ATF4 [Hs00909569_g1 (FAM)], and RPLP0 [Hs99999902_m1 (VIC)].

Western blot analysis

Cells were lysed in 1% LDS buffer (10 mmol/L Tris-HCl pH 8.0, 1% lithium dodecyl sulfate, 10% glycerol) and the protein concentration was quantified with detergent-compatible protein assay (Bio-Rad). Lysates (10–20 μg per lane) were resolved by SDS-PAGE and transferred onto PVDF membranes. After 1 hour of incubation with 5% skim milk in TBS-T or Can Get Signal/PVDF Blocking Reagent Set (NKB101/NYPBR; TOYOBO), membrane was incubated with primary antibodies diluted in 1% skim milk in TBS-T at 4°C overnight. The membrane was washed with TBS-T three times, then incubated with secondary antibodies diluted in 1% skim milk in TBS-T at room temperature for 1 hour. The membrane was again washed with TBS-T three times, then incubated with Luminata Forte Western HRP Substrate (WBLUF0500; Merck Millipore) and the signal was detected with ImageQuant LAS4000 imager (GE Healthcare). Antibodies used were anti-PTDSS2 [CEC clone 401-P-17, Lot. No. 18E07-IH (NH)], anti-β-actin (sc-69879; Santa Cruz Biotechnology), anti-FACL4 [EPR17587–42] (ab205119; Abcam), anti-CHOP (DDIT3) (2895; Cell Signaling Technology), anti-BIP (3177; Cell Signaling Technology), anti-cleaved PARP (9541; Cell Signaling Technology), HRP-linked anti-rabbit IgG antibody (NA934V; GE Healthcare), and HRP-linked anti-mouse IgG antibody (NA931V; GE Healthcare).

PTDSS1/2 protein preparation for compound screening

PTDSS1- or PTDSS2-expressing Sf9 cell membrane fractions were produced and isolated as follows. Flag-PTDSS1-HA or Flag-PTDSS2-HA was amplified in accordance with a previously described method (18) by PCR and cloned into pFastBac1 vector (10359016; Thermo Fisher Scientific). The vectors were transformed into DH10bac (10361012; Thermo Fisher Scientific) to prepare a bacmid. A PTDSS1- or PTDSS2-expressing baculovirus was prepared with the bacmids and infected into Sf9 cells (11496-015; Thermo Fisher Scientific). The Sf9 cells were recovered, suspended in buffer A [0.25 M sucrose, 10 mmol/L HEPES (pH 7.5), 1 mmol/L EDTA, 1 tablet/50 mL cOmplete EDTA-free (4693132; Roche Diagnostics)], fractured by ultrasonication, and then centrifuged (1,000 × g, 10 minutes, 4°C). The supernatant was recovered and treated by ultracentrifugation (100,000 × g, 1 hour, 4°C), and then the pellet was suspended in buffer A again. After ultracentrifugation (100,000 × g, 1 hour, 4°C) again, the pellet was suspended in harvest buffer [10 mmol/L HEPES (pH 7.5), 20% (v/v) glycerol, 1 tablet/100 mL cOmplete EDTA-free] to give a PTDSS1- or PTDSS2-expressing membrane fraction.

Compounds

DS07551382 (molecular formula C27H18ClF3N4O4, molecular weight 554.90) and DS55980254 (molecular formula C29H18F8N4O4, molecular weight 638.46) were prepared at Daiichi Sankyo Co., Ltd., as described previously (19). Thapsigargin was purchased from Sigma-Aldrich Co. LLC (catalog no. T9033). Tunicamycin was purchased from Wako Pure Chemical Industries, Ltd. (catalog no. 202-08241). For in vitro experiments, the compounds were dissolved in dimethyl sulfoxide (DMSO). DMSO was used for the control treatment.

Cell-free PS synthesis pulse chase assay

One hundred microliters of reaction solutions {50 mmol/L Hepes-NaOH (pH 7.5), 5 mmol/L CaCl2, 1 μCi/ml L-[14C(U)]-serine (NEC286E, PerkinElmer; or MC-265, Moravec), 0.8 mg/mL PTDSS1- or PTDSS2-expressing Sf9 cell membrane fraction} containing different concentrations of compounds were added to each well of 96-well plates and incubated at 37°C for 20 minutes. The reaction was stopped by adding 100 μL of 10 mmol/L EDTA, and the Sf9 cell membrane fractions in the reaction mixture were recovered using Unifilter-96 GF/C (6005174; PerkinElmer) and Unifilter Harvester (196; Packard). The filter was washed with 50 mmol/L Hepes-NaOH (pH 7.5) five times, air-dried, and then 40 μL of Microscint20 (6013621; PerkinElmer) was added dropwise to each well. Scintillation counts were measured on a TopCount-NXT-HTS (C384V01; PerkinElmer).

Cell-based PS synthesis pulse chase assay

For the evaluation of PTDSS1 siRNA-transfected cells, these cells were seeded at 4 × 104 cells/well onto a 96-well plate. The next day, the medium was replaced with minimum essential medium supplemented with 10% dialyzed bovine serum containing 5 μCi/ml L-[U-14C]-serine (NEC286E; PerkinElmer), and the cells were cultured for an additional 30 minutes. The cells were washed with PBS, after which, methanol was added to each well and incubated for 30 minutes at room temperature. The methanol was recovered and mixed with methyl tert-butyl ether, and the mixture was left to stand for 5 minutes. Distilled water was further added to the mixture and incubated for 10 minutes. The mixed solution was centrifuged at 3,000 rpm for 10 minutes, and the organic layer was collected and mixed with Pico-Fluor Plus (6013699; PerkinElmer). The radioactivity was measured using a liquid scintillation counter. The cell viability was measured with ATPlite 1step (6016739; PerkinElmer) and the scintillation counts were normalized with the ATP counts.

To evaluate the compound's inhibitory activity on PS synthesis in cells, HCT116 cells and HCT116 PTDSS2-KO#3 cells were prepared in McCoy's 5A medium containing 10% FBS at 20,000 cells/50 μL/well, and then seeded into 96-well plates and cultured at 37°C with 5% CO2 overnight. The medium was removed, 100 μL of MEM containing 10% bovine dialyzed serum with compound solutions (final concentration of dimethyl sulfoxide: 0.2%) and L-[14C(U)]-serine (final concentration: 2.5 μCi/mL) was added, followed by culture at 37°C with 5% CO2 for 24 hours. After cells had been washed once with PBS, 100 μL of methanol was added to each well, followed by incubation at room temperature for 30 minutes. Methanol was recovered in 96-well cluster tubes (4411; Corning), and 50 μL of chloroform and 50 μL of 50 mmol/L HEPES were added to each tube, which was stirred with a vortex mixer and then incubated at room temperature for 10 minutes. Fifty microliters of chloroform and 50 μL of 50 mmol/L HEPES were added to each tube, which was stirred with a vortex mixer and then centrifuged (240 × g, 4°C, 5 minutes). Ninety microliters of the organic layer of each tube was recovered and added to PicoPlate-96 (6005162; PerkinElmer). After the plates had been air-dried, 100 μL of Microscint-20 (6013621; PerkinElmer) was added to each well and scintillation counts were measured using TopCount-NXT-HTS.

Evaluation of compound activity on cell growth

The compounds were diluted and prepared using a Freedom EVO 150 (Tecan Trading; 4-fold dilution, 10 steps, 10 mmol/L–38 nmol/L) and were added to each well of 384-well plates at 100 nL/well using an Echo555 (Labcyte Inc.). HCT116, HCT116 PTDSS2-KO#3, and ARPE-19 cells were seeded into the plates at 200 cells (for HCT116 or HCT116 PTDSS2-KO#3) or 400 cells (for ARPE-19 cells)/40 μL/well (day 0) and cultured for 3 days. On the day of compound addition (day 0) and 3 days later (day 3), 10 μL/well of CellTiter-Glo(R) 2.0 Assay (G9242; Promega) was added to each well, and the amount of luminescence in each well was measured with EnVision. From the amount of luminescence on the day of compound addition (C0) and the amounts of luminescence of the compound non-addition group (C3) and compound addition group (T3) after 3 days of culture, cell viability was calculated using the following formula.

formula

Pharmacokinetic/pharmacodynamic analysis

HCT116 PTDSS2-KO#3 cells were transplanted subcutaneously into the right axillary region of female CAnN.Cg-Foxn1[nu]/CrlCrlj [Foxn1nu/Foxn1nu] mice (purchased from Charles River Japan) at a rate of 1 × 107 cells/head. When the tumor volume reached 100 to 300 mm3 [estimated tumor volume (major axis × minor axis × minor axis /2)], the mice were randomized, grouped with matching for tumor volume, and orally administered DS55980254. The vehicle control group was administered 0.5% methyl cellulose (MC). All animals were intraperitoneally administered 300 mg/kg deuterium (2H)-labeled serine [2,3,3-D3-L-serine, (Cambridge Isotope Laboratories, Inc., catalog no. DLM-582–0.5)] diluted in saline 1 hour before sample collection. Tumor samples were collected 24 hours after compound administration and cut into more than three pieces for subsequent analyses [pharmacokinetic (PK), pharmacodynamic (PD), and RNA-seq].

The lipid fraction of each tumor sample was extracted based on the Bligh & Dyer method. Briefly, samples were homogenized in methanol (20 μL/mg tumor) followed by centrifugation at 20,400 × g for 10 minutes. Four hundred microliters of the supernatant, which corresponds to the extract of 20 mg of tumor, from each sample was mixed with 200 μL of water and 400 μL of chloroform by a vortex mixer. The organic phase (400 μL) was collected and dried, and then the pellets were dissolved in 250 μL of 1% Triton X-100. To examine the amount of 2H-labeled serine incorporated into lipids, the extracted lipid fractions were mixed with 250 μL of phospholipase reaction mix [600 U/mL phospholipase D (Enzo Life Sciences, Inc., catalog no. BML-SE3011–0025), 50 mmol/L NaCl, 50 mmol/L Tris-HCl (pH 7.4)] and incubated at 37°C for 1 hour. Deuterium-labeled serine cleaved from the lipid fraction was extracted by collecting 500 μL of the aqueous phase from the mixture of phospholipase-treated samples (500 μL in total), 1 mL of methanol containing 5 μmol/L L-serine (13C, D3, 15N; Cambridge Isotope Laboratories, Inc., catalog no. CDNLM-6813–0.25) and 1 mL of chloroform. The concentration of 2H-labeled serine in samples was determined using LC/MS-MS. The ionization was conducted in positive ion mode using the transition m/z 109.2 to m/z 63.1. De novo PS synthesis in the tumor was calculated based on the following formula:

De novo PS synthesis (pmol/mg tumor/hour) = Observed 2H-labeled serine conc. (μmol/L) × Sample volume (μL) / 20 (mg of tumor)

Tumor homogenate for PK measurement was prepared by homogenizing tumor samples in distilled water (0.45 mL/50 mg of tumor). The concentration of DS55980254 in tumor homogenate was determined using LC/MS-MS. The ionization was conducted in positive ion mode using the transition m/z 639.1 to m/z 619.1.

RNA was isolated from tumor samples using Maxwell RSC SimplyRNA Tissue Kit (Promega Corporation, catalog no. AS1340) and used for subsequent RNA-seq analysis.

Evaluation of in vivo efficacy of DS55980254

HCT116 wild-type, HCT116 PTDSS2-KO#3 or A375 PTDSS2-KO#54 cells were transplanted subcutaneously into the right axillary region of female CAnN.Cg-Foxn1[nu]/CrlCrlj [Foxn1nu/Foxn1nu] mice (purchased from Charles River Japan) at a rate of 1 × 107 (HCT116) or 1 × 106 (A375) cells/head. When the tumor volume reached 100 to 300 mm3 [estimated tumor volume (major axis × minor axis × minor axis/2)], the mice were randomized and grouped to ensure matching of tumor volume among the groups. From the day of grouping, DS55980254 was orally administered once a day for 14 days for wild-type HCT116, for 28 days for HCT116 PTDSS2-KO#3 and for 11 days for A375 PTDSS2-KO#54. The tumor volume and body weight were measured until the day after the completion of administration. The vehicle control group was administered 0.5% MC. The body weight change was calculated from the body weight on the first day of administration (W0) and the body weight on each day (W) using the following formula.

formula

Nonclinical safety

To investigate the effect of DS55980254 on human ether-a-go-go related gene (hERG) potassium current, manual patch clamp assay was conducted using a recombinant CHO-K1 cell line introduced with hERG. The cells for measurement were seeded on 0.1% gelatin-coated glass. DS55980254 dissolved in Tyrode solution was directly exposed to the cells with the manifold injection system (flow rate: 1 mL/min, room temperature). The whole-cell patch clamp method was applied using electrodes with resistance of 1 to 4 mΩ. The membrane potential was maintained at −80 mV, followed by a depolarization pulse for 2 seconds at 40 mV. Then, a repolarization pulse for 2 seconds at −50 mV was applied at 0.067 Hz. Data acquisition and analyses were performed using a commercial patch clamp amplifier and pCLAMP program (MDS-AT).

To investigate the proarrhythmic potential of DS55980254, the rabbit isolated Langendorff heart model was used. The heart isolated from female NZW rabbit (Oriental Yeast Co., Ltd.) was retrogradely perfused at the coronary artery at a constant flow rate of about 24 mL/min with Krebs–Henseleit buffer. Two ECG electrodes were held lightly against the epicardium, the positive one on the left ventricle, the negative one on the electro cannula. A monophasic action potential (MAP) recording electrode was attached to the wall of the left ventricle to obtain an epicardial MAP signal. Increasing concentrations of DS55980254 dissolved in the perfusate were exposed to the heart. QT and QRS intervals, and MAP duration were recorded before and 20 minutes after exposure of the test compound at each concentration.

To investigate the general toxicity profile of DS55980254, 14-day repeated oral dose toxicity study was conducted in mice. Six-week-old male Crl:CD1(ICR) mice were purchased from Charles River Japan and then housed for 1 week. Mice (n = 5) were orally dosed once daily by gavage with vehicle or DS55980254 at 100, 300, and 1000 mg/kg in 0.5% methylcellulose for 14 days. Animals were euthanized the day after the last dose, and their tissues and organs were collected for organ weight assessment and histopathological examination. Blood samples were obtained from the abdominal aorta and subjected to hematology and serum chemistry assays.

RNA-seq

HCT116, A375, HCT116 PTDSS2-KO#3, and A375 PTDSS2-KO#54 cells were seeded at 1 × 105 cells/well onto a 12-well plate. DMSO or DS07551382 (final concentration: 0.1 μmol/L) was added to each well and incubated for 24 hours. RNA was isolated using RNeasy Mini Kit (74106; Qiagen). For in vitro samples, a cDNA library for RNA-seq was prepared using the Poly(A) enrichment method of mRNA with NEBNext Poly(A) mRNA Magnetic Isolation Module (E7490; New England Biolabs) and NEBNext Ultra RNA Library Prep Kit for Illumina (E7530; New England Biolabs), following the manufacturer's protocol. Then, 250 ng of total RNA was used as input for the preparation. The prepared libraries were evaluated for quality and quantity using 4200 TapeStation system. In addition, qPCR was carried out using KAPA Library Quantification Kit Illumina Platforms (KAPA Biosciences). For in vivo samples, a cDNA library was prepared using SMARTer Stranded Total RNA-Seq Kit v2-Pico Input Mammalian (634413; Takara Bio) in accordance with the manufacturer's protocol. Five nanograms of total RNA was used as an input for the preparation. The prepared libraries were evaluated in terms of quality and quantity using the D1000 Assay Kit of the 4200 TapeStation system.

The prepared libraries were pooled and sequenced on NextSeq500 (Illumina) with 75 bp from a single end (1 × 75 bp). The sequencing reads were aligned with STAR software (2.4.2a; ref. 20) to the transcript reference, hg19 (human reference) for in vitro samples, and both hg19 (human reference) and mm10 (mouse reference) for in vivo samples. For in vivo samples, human- and mouse-specific reads were defined as follows:

Human-specific reads = Human mapped reads with 0.04 mismatch rate − (Mouse mapped reads with 0.02 mismatch rate + Human and mouse mapped reads with 0.04 and 0.02 mismatched rates)

Mouse-specific reads = Mouse mapped reads with 0.04 mismatch rate − (Human mapped reads with 0.02 mismatch rate + Mouse and human mapped reads with 0.04 and 0.02 mismatched rates)

Transcripts per million (TPM), count, and fragments per kilobase of exon per million reads mapped (FPKM) in each gene were estimated by RSEM software (1.2.23; ref. 21). The R package DESeq2 was used for data normalization and differential expression analysis. Pathway enrichment analysis was performed using GSEA software based on the normalized counts by DESeq2. GSEA was implemented with Java Web Start version 2.2.1 from the Broad Institute (22).

Human mobility group box 1 detection assay

Human mobility group box 1 (HMGB1) detection was performed using HMGB1 ELISA Kit II (326054329; Shino Test) following the manufacturer's protocol. HCT116 PTDSS2-KO#3, A375 PTDSS2-KO#54, and CT26.WT Pss2-KO#1 cells were seeded at 2 × 104, 1 × 104, or 1 × 104 cells/well onto a 96-well plate, respectively. The cells were treated with DMSO or DS55980254 for 48 hours and the conditioned media were subjected to HMGB1 detection.

In vitro dendritic cell differentiation assay

Mouse bone marrow-derived dendritic cells (BMDC) were isolated based on a previously reported method (23–25). Bone marrow cells isolated from female BALB/cAnNCrlCrlj mice (purchased from Charles River Japan) were cultured for 7 days in RPMI1640 medium containing 10% FBS, 10 mmol/L Hepes, 1 mmol/L sodium pyruvate, 100 units/mL penicillin, 100 μg/mL streptomycin, 55 μmol/L 2-mercaptoethanol, and 20 ng/mL GM-CSF. Nonadherent cells were collected and used for subsequent experiments.

CT26.WT or CT26.WT Pss2-KO#1 cells were seeded at 1 × 106 cells onto a 10 cm dish. The following day, cells were treated with the indicated concentration of DMSO or DS55980254 for 24 hours. The cells were recovered, resuspended in DMSO- or DS55980254-containing medium at 1 × 106 cells/mL, and seeded onto a U-bottomed 96-well plate (100 μL/well). Mouse BMDCs derived from BALB/cAnNCrlCrlj mice were added to each well at 2 × 104 cells/well and further incubated overnight. Mean fluorescence intensity of CD80 or CD86 positivity in 7-aminoactinomycin D (7-AAD)-negative and CD11b/CD11c double-positive cells was quantified by MACSQuant X. The reagents and antibodies used were 7-AAD (51-68981E; BD Pharmingen), PE Rat Anti-CD11b Antibody (557397; BD Pharmingen), APC anti-mouse CD11c Antibody (117310; BioLegend), APC/Fire750 anti-mouse CD80 Antibody (104739; BioLegend), and Brilliant Violet 421 anti-mouse CD86 Antibody (105031; BioLegend). Data were analyzed using FlowJo version 7.6.5.

In vivo antitumor test for syngeneic PTDSS2-intact/null heterogeneity model

CT26.WT and CT26.WT Pss2-KO#1 cells were mixed at a ratio of 5:95 and transplanted subcutaneously into the right axillary region of female BALB/cAJcl-nu/nu or BALB/cAJcl mice (purchased from CLEA Japan) at a rate of 3 × 105 cells/head. As a control, groups transplanted with unmixed CT26.WT Pss2-KO#1 cells under the same conditions were prepared. From the day after inoculation, mice were randomized as five per group and 100 mg/kg DS55980254 was orally administered once a day for 13 days. The estimated tumor volume (major axis×minor axis/2) and body weight were measured until the day after the completion of administration. The vehicle control group was administered 0.5% MC. The T-cell–mediated effect on tumor was calculated by comparing the mean tumor volumes of nude mice in each treatment with the tumor volume in Balb/c mice of the same treatment group on the final day. Namely, the tumor volume of each Balb/c mouse (vol_Balb) in each group (vehicle or DS55980254) was normalized with the mean volume of tumor in nude mice with the same treatment (M), and the T–cell-mediated antitumor effect was calculated using the following formula.

formula

Statistical analysis

GraphPad Prism 9 software was used to visualize and analyze the data. Two-way ANOVA was used to calculate P values for siRNA knockdown experiments (Fig. 1E and F). One-way ANOVA was used to calculate P values for the comparisons of DDIT3 and ATF4 expression (Fig. 3E), HMGB1 detection assay (Fig. 4B), and in vitro dendritic cell differentiation assay (Fig. 4D). Unpaired t test was used to calculate the P value of the T-cell-mediated effect in in vivo analysis (Fig. 4F).

Figure 1.

Loss of PTDSS2 confers vulnerability to PTDSS1 depletion. A, Integrative Genomics Viewer image showing chromosome 11p15.5 locus. B, Bar plot showing the percentage of patients with deep deletion of the indicated genes in the indicated cancer types. C, Representative images of PTDSS2 staining for patients with non–small cell lung cancer in each category. S, stroma; T, tumor. Scale bar, 50 μm. D, Bar graph showing percentages of patients in each category in each cancer type. The number shows the patient number in each category. TNBC, triple-negative breast cancer; NSCLC, non–small cell lung cancer; AC, adenocarcinoma; TCC, transitional cell cancer; CA, carcinoma; SCC, squamous cell carcinoma; H&N, head and neck; LCC, large cell carcinoma; ER, estrogen receptor. E, Growth curves showing the specific vulnerability of PTDSS2-knockout cell clones to PTDSS1 depletion. The mean and SD of four replicates are indicated. The results of statistical analysis on day 8: ****, P < 0.0001. ns, not significant. F, Bar graph showing the incorporation of 14C-serine into the membrane fraction in each group. The mean ± SD of triplicate are indicated. **, P < 0.01; ****, P < 0.0001.

Figure 1.

Loss of PTDSS2 confers vulnerability to PTDSS1 depletion. A, Integrative Genomics Viewer image showing chromosome 11p15.5 locus. B, Bar plot showing the percentage of patients with deep deletion of the indicated genes in the indicated cancer types. C, Representative images of PTDSS2 staining for patients with non–small cell lung cancer in each category. S, stroma; T, tumor. Scale bar, 50 μm. D, Bar graph showing percentages of patients in each category in each cancer type. The number shows the patient number in each category. TNBC, triple-negative breast cancer; NSCLC, non–small cell lung cancer; AC, adenocarcinoma; TCC, transitional cell cancer; CA, carcinoma; SCC, squamous cell carcinoma; H&N, head and neck; LCC, large cell carcinoma; ER, estrogen receptor. E, Growth curves showing the specific vulnerability of PTDSS2-knockout cell clones to PTDSS1 depletion. The mean and SD of four replicates are indicated. The results of statistical analysis on day 8: ****, P < 0.0001. ns, not significant. F, Bar graph showing the incorporation of 14C-serine into the membrane fraction in each group. The mean ± SD of triplicate are indicated. **, P < 0.01; ****, P < 0.0001.

Close modal

Data availability

Gene alteration profiles (deep deletion) of PTDSS2, HRAS, RASSF7, and CDKN1C were obtained from cbioportal.org. RNA-seq data supporting the findings of this study have been deposited in Gene Expression Omnibus (GEO) with accession code GSE206518.

Synthetic lethal relationship of PTDSS1/2 in human cancer cells

As PTDSS2 is located at a tumor-suppressive locus, 11p15.5 (Fig. 1A), we first investigated whether PTDSS2 is lost in tumors using clinical samples. TCGA dataset analysis showed that PTDSS2 is homozygously deleted in a wide range of cancer types together with genes at 11p15.5 such as HRAS, RASSF7, and CDKN1C (Fig. 1B; Supplementary Fig. S1). To further confirm the loss of PTDSS2 protein expression in cancer patients, we raised a PTDSS2 antibody and confirmed its specificity using PTDSS2-knockout (KO) cell lines (Supplementary Fig. S2). The genomic deletion of PTDSS2 in HCT116 PTDSS2-KO#3, #16, and A375 PTDSS2-KO#54 was confirmed by genomic PCR (Supplementary Figs. S2A and S2B). We observed heterozygous PTDSS2 deletion in A375 PTDSS2-het#27 clone (Supplementary Fig. S2B). Western blotting and IHC using these cell lines showed strong staining only in PTDSS2 wild type and weak staining in heterozygously deleted clones, suggesting the high specificity of the antibody (Supplementary Figs. S2C and S2D). Using this antibody, we conducted IHC analysis with TMAs of tumors from cancer patients. On the basis of the staining intensity and pattern, we evaluated the stained TMAs and interpreted the status using the following four categories (Fig. 1C): “intact,” having weakly to strongly positive staining in the entire/most of the tumor; “homogeneous loss,” showing no staining in the entire/most of the tumor with or without positive staining of adjacent non-neoplastic stromal tissue; “heterogeneous loss,” showing a tumor with positive and negative staining (>50%); and “indeterminate,” exhibiting faint/not clear-cut staining with or without positive staining of adjacent stromal tissue. PTDSS2 protein loss was observed in a wide range of cancer types (Fig. 1D). Next, we investigated whether the synthetic lethality between PTDSS1 and PTDSS2 previously observed in mice (9) is conserved in human cancer cells. Although a wide spectrum of PTDSS2 deletions was observed in the patients, no PTDSS2-null cell line was conventionally available. Thus, we utilized PTDSS2-KO HCT116 cell lines for validation. PTDSS1 depletion mediated by siRNA (Supplementary Fig. S2E) inhibited the growth of PTDSS2-KO HCT116 clones, whereas it did not affect that of wild-type HCT116 cells (Fig.1E). In line with this observation, de novo PS production, which was monitored by the incorporation of 14C-labeled serine into the cellular membrane fraction, was significantly suppressed upon PTDSS1/2 dual depletion compared with that upon single PTDSS1 depletion (Fig. 1F). These results suggest that the loss of PTDSS2 confers specific vulnerability in PS production, which could be targeted by PTDSS1 inhibition.

Development of a potent and selective inhibitor of PTDSS1

To obtain small-molecule inhibitors against the PS synthetase activity of PTDSS1, we first established a cell-free assay system that detects PTDSS activity. Human PTDSS1 and PTDSS2 were expressed in Sf9 insect cells using a baculoviral expression system and the microsome membrane fraction was purified. Using this membrane fraction as an enzyme source, we conducted chemical screening and derivatization and obtained pyrrolepyrazole derivatives, DS07551382 and DS55980254 (Fig. 2A). DS07551382 and DS55980254 suppressed the PS production activity of PTDSS1 but not that of PTDSS2 in the cell-free assay, suggesting that the inhibitors are selective to PTDSS1 over PTDSS2 (Fig. 2B). DS07551382 and DS55980254 strongly suppressed the de novo PS synthesis in PTDSS2-KO HCT116 clone, whereas they weakly suppressed it in parental HCT116 (Fig. 2C; Supplementary Fig. S3A). In addition, these compounds selectively suppressed the growth of PTDSS2-KO HCT116 in vitro, which differed strikingly compared with PTDSS2-wild-type HCT116 or ARPE-19, a nonmalignant pigment retinal epithelial cell line (Fig. 2D; Supplementary Fig. S3B). As DS55980254 was highly bioavailable, we used it for in vivo study. We observed the decrease of de novo PS synthesis in the tumor, which was monitored by determining the incorporation of deuterium-labeled serine into the lipid fraction, in an intratumor DS55980254-concentration-dependent manner, confirming the pharmacokinetic/pharmacodynamic relationship (Supplementary Fig. S3C). We observed the tumor regression in PTDSS2-KO HCT116 and PTDSS2-KO A375 subcutaneous xenograft models upon treatment with a low dose (10–30 mg/kg) of DS55980254 without body weight loss (Fig. 2E and F). DS55980254 treatment did not suppress the growth of parental HCT116 xenograft (Supplementary Fig. S3D). Preclinical safety analysis showed that DS55980254 does not have a potential risk for hERG current inhibition and proarrhythmic effects in the rabbit Langendorff model at concentrations up to 30 μmol/L, and does not induce severe toxicity in mice after 14-day repeated oral administration up to 1,000 mg/kg/day (Table 1). These results show that these pyrrolepyrazole derivatives are highly specific inhibitors against PTDSS1 and can induce selective cell death in PTDSS2-deleted cancer cells without severe toxicity.

Figure 2.

The development of PTDSS1 selective inhibitors. A, Chemical structures of DS07551382 and DS55980254. B, Cell-free inhibitory activity of DS07551382 and DS55980254 against PTDSS1 and PTDSS2. Means ± SD of four replicates (for PTDSS1) or three replicates (for PTDSS2) are shown. C, The inhibitory activity of DS55980254 on PS production in parental or PTDSS2-KO HCT116. Means ± SD of three replicates are shown. D, The effect of DS55980254 on cell growth. Means ± SD of four replicates are shown. E and F, Tumor-bearing mice were treated with DS55980254 (10, 30, 100 mg/kg, daily). The tumor volume and body weight change are plotted. Means ± SEM are shown (N = 6). E, HCT116 PTDSS2-KO#3; F, A375 PTDSS2-KO#54.

Figure 2.

The development of PTDSS1 selective inhibitors. A, Chemical structures of DS07551382 and DS55980254. B, Cell-free inhibitory activity of DS07551382 and DS55980254 against PTDSS1 and PTDSS2. Means ± SD of four replicates (for PTDSS1) or three replicates (for PTDSS2) are shown. C, The inhibitory activity of DS55980254 on PS production in parental or PTDSS2-KO HCT116. Means ± SD of three replicates are shown. D, The effect of DS55980254 on cell growth. Means ± SD of four replicates are shown. E and F, Tumor-bearing mice were treated with DS55980254 (10, 30, 100 mg/kg, daily). The tumor volume and body weight change are plotted. Means ± SEM are shown (N = 6). E, HCT116 PTDSS2-KO#3; F, A375 PTDSS2-KO#54.

Close modal
Table 1

Nonclinical safety profile of DS55980254.

StudyResults
hERG current inhibition IC50 > 30 μmol/L 
Rabbit isolated Langendorff heart modela No effect up to 30 μmol/L 
Mouse 14-day repeated oral dose toxicity studyb (0, 100, 300, 1,000 mg/kg, every day dosing) Target organs 
 Liver: 
 - Fatty change in centrilobular hepatocytes at 1,000 mg/kg 
 Hardarian gland: 
 - Inflammatory cell infiltration and regeneration in acinar cells at 100 mg/kg and more. 
StudyResults
hERG current inhibition IC50 > 30 μmol/L 
Rabbit isolated Langendorff heart modela No effect up to 30 μmol/L 
Mouse 14-day repeated oral dose toxicity studyb (0, 100, 300, 1,000 mg/kg, every day dosing) Target organs 
 Liver: 
 - Fatty change in centrilobular hepatocytes at 1,000 mg/kg 
 Hardarian gland: 
 - Inflammatory cell infiltration and regeneration in acinar cells at 100 mg/kg and more. 

aThe following cardiac parameters were evaluated in this study: QT interval, QRS duration, and MAPD90.

bThe following examination items were evaluated in this study: clinical signs, body weight, food consumption, hematology, blood chemistry, organ weight, histopathology, and toxicokinetics.

Inhibition of PS synthesis induces ER stress

To decipher the underlying mechanism governing selective growth arrest in PTDSS1-inhibited PTDSS2-depleted cancer cells, we analyzed the transcriptomic changes brought about by comparing DMSO-treated and DS07551382-treated wild-type or PTDSS2-depleted cells. Gene set enrichment analysis (GSEA) showed that no signature was significantly enriched by DS07551382 treatment in HCT116 parental cells, whereas numerous signatures were enriched in HCT116 PTDSS2-KO#3 cells (Supplementary Fig. S4A, left). Similarly, many more signatures were enriched by DS07551382 treatment in A375 PTDSS2-KO#54 cells than in A375 parental cells (Supplementary Fig. S4A, right). The hallmark “unfolded protein response,” which contains genes related to ER stress response (26), was the most enriched signature commonly upregulated in DS07551382-treated PTDSS2-depleted cells (Fig. 3A and B). Enrichment plots showed that the enrichment of this signature is specific to PTDSS2-depleted cells (Fig. 3C and D). The selective upregulation of ER stress-related genes, such as DDIT3 and ATF4, in PTDSS1-inhibited PTDSS2-KO cells was further confirmed using another PTDSS1 inhibitor, DS55980254 (Fig. 3E). The level of induction of expression of these genes was comparable to that of classical ER stress-inducing agents, such as thapsigargin and tunicamycin (Supplementary Fig. S4B). The ER stress induction was also confirmed by Western blotting analysis, which revealed the upregulation of BIP and DDIT3 protein in DS55980254-treated HCT116 PTDSS2-KO#3 cells (Fig. 3F). A significant increase of cleaved PARP was observed only in DS55980254-treated HCT116 PTDSS2-KO#3 cells (Fig. 3F). Whole-transcriptome analysis in vivo comparing vehicle- and DS55980254-treated PTDSS2-KO HCT116 xenografts also showed the upregulation of ER stress upon DS55980254 treatment (Fig. 3G and H). These results suggest that PTDSS1 inhibitor induces ER stress downstream of PS depletion in PTDSS2-deleted tumor cells.

Figure 3.

Selective enrichment of unfolded protein response signature by PTDSS1 inhibitor in PTDSS2-deleted cancer cells. A and B, GSEA (hallmark) showing enriched signatures by PTDSS1 inhibition comparing DMSO- and DS07551382-treated HCT116 PTDSS2-KO#3 (A) or A375 PTDSS2-KO#54 (B). Signatures with P < 0.05 and FDR q value < 0.25 are shown. C and D, Enrichment plot showing hallmark unfolded protein response upon DS07551382 treatment in HCT116 (C) and A375 (D) parental or PTDSS2-KO cell line. E, Bar graphs showing the expression of DDIT3 and ATF4 in HCT116 and HCT116 PTDSS2-KO#3 cells treated with DMSO or DS55980254 for 48 hours. Means ± SD of three replicates are shown. ****, P < 0.0001. F, Western blot analysis showing the selective upregulation of DDIT3, BIP, and cleaved PARP in DS55980254-treated HCT116 PTDSS2-KO#3 clone. G, Enrichment plot showing hallmark unfolded protein response in PTDSS2-KO HCT116 xenograft comparing vehicle vs. DS55980254 treatment in vivo. H, Plots showing the expression of ER stress-related genes in vehicle- and DS55980254-treated PTDSS2-KO HCT116 xenografts. The individual and median normalized counts of each group are shown (N = 6).

Figure 3.

Selective enrichment of unfolded protein response signature by PTDSS1 inhibitor in PTDSS2-deleted cancer cells. A and B, GSEA (hallmark) showing enriched signatures by PTDSS1 inhibition comparing DMSO- and DS07551382-treated HCT116 PTDSS2-KO#3 (A) or A375 PTDSS2-KO#54 (B). Signatures with P < 0.05 and FDR q value < 0.25 are shown. C and D, Enrichment plot showing hallmark unfolded protein response upon DS07551382 treatment in HCT116 (C) and A375 (D) parental or PTDSS2-KO cell line. E, Bar graphs showing the expression of DDIT3 and ATF4 in HCT116 and HCT116 PTDSS2-KO#3 cells treated with DMSO or DS55980254 for 48 hours. Means ± SD of three replicates are shown. ****, P < 0.0001. F, Western blot analysis showing the selective upregulation of DDIT3, BIP, and cleaved PARP in DS55980254-treated HCT116 PTDSS2-KO#3 clone. G, Enrichment plot showing hallmark unfolded protein response in PTDSS2-KO HCT116 xenograft comparing vehicle vs. DS55980254 treatment in vivo. H, Plots showing the expression of ER stress-related genes in vehicle- and DS55980254-treated PTDSS2-KO HCT116 xenografts. The individual and median normalized counts of each group are shown (N = 6).

Close modal

Immunogenic cell death induced by PTDSS1 inhibitor activates tumor immunity

The existence of intratumor heterogeneity of PTDSS2 loss (Fig. 1C and D) raised concerns that residing PTDSS2-intact tumor cells escape PTDSS1 inhibitor treatment and potentially cause recurrence or resistance. Previous studies demonstrated that ER stress-induced cell death can be immunogenic through the secretion of DAMPs, such as high HMGB1 protein (27, 28). We observed that PTDSS1 inhibitor induces selective cell death in PTDSS2-deleted tumor accompanied by the induction of ER stress (Fig. 3). These findings led us to hypothesize that PTDSS1 inhibitor can induce immunogenic cell death in PTDSS2-deleted tumor, thereby activating tumor immunity and leading to surrounding PTDSS2-intact tumor cells being attacked by activated immune cells (Fig. 4A).

Figure 4.

Activation of antitumor immunity by immunogenic cell death of PTDSS2-deleted tumor cells by PTDSS1 inhibitor. A, Schematic diagram showing the proposed model of the antitumor activity on PTDSS2 wild-type cells through the activation of tumor immunity by inducing immunogenic cell death in PTDSS2-null tumor cells. DC, dendritic cell. Created with BioRender.com. B, Bar graphs showing the level of HMGB1 in culture conditioned medium after DS55980254 treatment in PTDSS2-KO cell lines. Means ± SD of three replicates are shown. C, A schematic diagram showing the flow of the experiment coculturing PTDSS1 inhibitor-treated tumor cells with dendritic cells. Created with BioRender.com. D, Bar graphs showing the mean fluorescent intensity (MFI) of mature dendritic cell markers on CD11b/CD11c double-positive cells for each treatment. Means ± SD of three replicates are shown. E, Tumor-bearing mice with the mixed model composed of parental and Pss2-KO#1 CT26.WT cells were treated with DS55980254 (100 mg/kg, daily). The tumor volume for each mouse is plotted as a spider plot (N = 5). F, A bar graph showing the T-cell–mediated antitumor effect. Means ± SD are shown (N = 5). ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure 4.

Activation of antitumor immunity by immunogenic cell death of PTDSS2-deleted tumor cells by PTDSS1 inhibitor. A, Schematic diagram showing the proposed model of the antitumor activity on PTDSS2 wild-type cells through the activation of tumor immunity by inducing immunogenic cell death in PTDSS2-null tumor cells. DC, dendritic cell. Created with BioRender.com. B, Bar graphs showing the level of HMGB1 in culture conditioned medium after DS55980254 treatment in PTDSS2-KO cell lines. Means ± SD of three replicates are shown. C, A schematic diagram showing the flow of the experiment coculturing PTDSS1 inhibitor-treated tumor cells with dendritic cells. Created with BioRender.com. D, Bar graphs showing the mean fluorescent intensity (MFI) of mature dendritic cell markers on CD11b/CD11c double-positive cells for each treatment. Means ± SD of three replicates are shown. E, Tumor-bearing mice with the mixed model composed of parental and Pss2-KO#1 CT26.WT cells were treated with DS55980254 (100 mg/kg, daily). The tumor volume for each mouse is plotted as a spider plot (N = 5). F, A bar graph showing the T-cell–mediated antitumor effect. Means ± SD are shown (N = 5). ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Close modal

To test this, we first established a Ptdss2 (Pss2)-KO mouse cancer cell line using the CRISPR/Cas9 system containing gRNAs targeting Exons 2, 4, and 5 (Supplementary Fig. S5A). As the PTDSS2 antibody that we raised was human-specific, we confirmed Pss2 gene deletion in CT26.WT cells by RT-PCR using a primer set covering Exons 1–6 in Pss2. A shift of around 200 bp was observed in CT26.WT Pss2-KO#1 clone (Supplementary Fig. S5B). Next, we examined the HMGB1 level in the culture conditioned medium of PTDSS1-inhibitor-treated PTDSS2-KO cells. The elevation of HMGB1 level was observed in DS55980254-treated PTDSS2-KO cell lines in a dose-dependent manner (Fig. 4B). To support the notion that DAMPs secreted from dying PTDSS2-deleted cells by PTDSS1 inhibitor can activate dendritic cells, we cultured BMDCs with PTDSS2-KO cells treated with PTDSS1 inhibitor in vitro (Fig. 4C). Coculture with DS55980254-treated CT26.WT Pss2-KO#1 increased the surface expression of CD80/CD86 in BMDCs in a DS55980254 dose-dependent manner, which was not observed in the coculture with DS55980254-treated parental CT26.WT (Fig. 4D). Finally, we examined the efficacy of PTDSS1 inhibitor on the PTDSS2-intact/null heterogeneity model in vivo. First, we confirmed that DS55980254 completely inhibited the growth of CT26.WT Pss2-KO#1 tumor in both BALB/c mice and the corresponding T-cell-deficient strain, BALB/c nude mice (Supplementary Fig. S5C). Next, we evaluated the efficacy of DS55980254 on CT26.WT-intact and Pss2-KO heterogeneity model. In nude mice, CT26.WT-intact tumor cells continued growing under DS55980254 treatment (Fig. 4E, nude); however, this growth was suppressed in BALB/c mice (Fig. 4E, Balb/c). The T-cell-mediated tumor-suppressing effect, which was calculated by comparing the antitumor effect between BALB/c mice and BALB/c nude mice within the same treatment, was greater in the DS55980254-treated group (Fig. 4F). These results suggest that PTDSS1 inhibitor may exhibit potent and durable antitumor efficacy by modulating the tumor immune microenvironment, even in an environment with heterogeneous PTDSS2 loss.

In this study, we identified PTDSS1 as a synthetic lethal target for tumors with PTDSS2 deletion combined with 11p15.5 loss. We showed that PTDSS1 depletion specifically suppresses the growth of PTDSS2-deleted cancer cells. We developed potent and selective PTDSS1 inhibitors that can eliminate PTDSS2-deleted tumor in vivo without severe toxicity. Whole-transcriptome analysis revealed that PS depletion mediated by PTDSS1 inhibitor upregulates ER stress in PTDSS2-deleted cells. Furthermore, we revealed that PTDSS1 induces immunogenic cell death in PTDSS2-deleted tumor cells, which in turn activates tumor immunity to suppress the growth of surrounding PTDSS2-intact tumor cells. Together, these results suggest that our PTDSS1 inhibitor provides a therapeutic option for patients with PTDSS2-deleted tumor, by inducing collateral lethality and activation of the tumor immune microenvironment (Fig. 5).

Figure 5.

Model showing PTDSS1 inhibitor's efficacy on tumor with PTDSS2 deletion. PTDSS1 inhibitor induces synthetic lethality to PTDSS2-deleted tumor cells, which involves immunogenic cell death through ER stress activation; this in turn leads to tumor immunity activation. DC, dendritic cell. Created with BioRender.com.

Figure 5.

Model showing PTDSS1 inhibitor's efficacy on tumor with PTDSS2 deletion. PTDSS1 inhibitor induces synthetic lethality to PTDSS2-deleted tumor cells, which involves immunogenic cell death through ER stress activation; this in turn leads to tumor immunity activation. DC, dendritic cell. Created with BioRender.com.

Close modal

Our analysis using clinical samples confirmed that PTDSS2 loss appeared together with potential tumor suppressor genes at the 11p15.5 locus (Fig. 1), suggesting that PTDSS2 deletion is collateral damage associated with 11p15.5 loss. The presence of a synthetic lethal relationship between PTDSS1 and PTDSS2 was reported in mice (9) and the importance of PTDSS1 as a target for PTDSS2-deleted cancer was predicted by computational analysis (29). We showed that siRNA-mediated PTDSS1 depletion specifically perturbs the growth of PTDSS2-deleted models, further supporting the synthetic lethal relationship between PTDSS1 and PTDSS2. Our observation that PS synthesis was not significantly affected by single depletion of PTDSS1 or PTDSS2 but strongly attenuated by their dual depletion in human cancer cells (Fig. 1F), which is also consistent with a report on Pss1/Pss2-knockout mice (9), supports the view that PTDSS1 and PTDSS2 play redundant roles in PS production in human cancer cells. Together, we conclude that PTDSS2 deletion is an attractive collateral lethal vulnerability created by 11p15.5 loss.

A key finding of this study is the discovery of potent, selective, and bioavailable PTDSS1 chemical inhibitors (Fig. 2). Our compounds strongly suppressed PS production activity and the growth of PTDSS2-deleted cancer cells, which clearly phenocopied the PTDSS1 inhibition mediated by siRNA. The fact that global change in gene signatures by DS07551382 treatment was predominantly observed in PTDSS2-deleted cells derived from two independent cell lines (Supplementary Fig. S4) also supports the compound's selective inhibitory activity on PTDSS1. The induction of ER stress in two independent PTDSS2-deleted cells by DS07551382 (Fig. 3A and B) indicates the fact that ER stress is one of the major pathways upregulated by the compound. This is consistent with previous reports describing that an imbalance of phospholipids upregulates ER stress signaling (30–32), and suggests that our compound induces PS imbalance in PTDSS2-deleted cells. Furthermore, DS55980254 did not show critical toxicity up to 1,000 mg/kg in mice (Table 1), which is consistent with the phenotype of Pss1 single-knockout mice (9), suggesting that the compound is highly specific to PTDSS1 and has almost no clear off-target effects.

As the deletion of PTDSS2 is likely to be a passenger event with 11p15.5 loss, the existence of intratumor heterogeneity of PTDSS2 loss is a natural consequence, considering the process of tumor evolution (33). In fact, we observed heterogeneous patterns of PTDSS2 staining in a wide range of cancer patients (Fig. 1C). Our compounds did not affect the growth of PTDSS2-intact cells in vitro or in vivo (Fig. 2D; Supplementary Fig. S3D), which raised the possibility that our compounds will not be effective on tumors with heterogeneous PTDSS2 loss. The finding that DS55980254 treatment increased the amount of HMGB1 in culture conditioned medium of PTDSS2-deleted cancer cells (Fig. 4B) suggests that PTDSS1 inhibitor-mediated cell death could enhance DAMPs in the tumor microenvironment. A coculture experiment with mouse BMDCs (Fig. 4D) further supported the notion that synthetic lethality induced by PTDSS1 inhibitor on PTDSS2-deleted tumors can activate dendritic cell maturation. The in vivo syngeneic heterogeneity model (Fig. 4E and F) showed that PTDSS1 inhibitor can suppress the growth of PTDSS2 intact cells through tumor immunity activated by the immunogenic cell death of PTDSS2 null cells. These results suggest that PTDSS1 inhibitor may overcome tumor heterogeneity in PTDSS2 loss by modulating the tumor immune microenvironment.

In conclusion, we present a potent and selective inhibitor of PTDSS1, which should provide a novel therapeutic option for cancer patients with PTDSS2 deletion as collateral damage of 11p15.5 loss.

Y. Yoshihama reports personal fees from Daiichi Sankyo Co. Ltd. during the conduct of the study and personal fees from Daiichi Sankyo Co. Ltd. outside the submitted work; also has a patent for WO2016148115 A1 pending and a patent for WO2020179859 A1 pending. H. Namiki reports personal fees from Daiichi Sankyo Co. Ltd. during the conduct of the study and also has a patent for WO2020179859 A1 pending. T. Kato reports personal fees from Daiichi Sankyo during the conduct of the study and personal fees from Daiichi Sankyo outside the submitted work. N. Shimazaki reports personal fees from Daiichi Sankyo during the conduct of the study and personal fees from Daiichi Sankyo outside the submitted work. S. Takaishi reports personal fees from Daiichi Sankyo during the conduct of the study and personal fees from Daiichi Sankyo outside the submitted work. T. Shibutani reports personal fees from Daiichi Sankyo during the conduct of the study and personal fees from Daiichi Sankyo outside the submitted work. M. Hirasawa reports personal fees from Daiichi Sankyo Co., Ltd. during the conduct of the study and personal fees from Daiichi Sankyo Co., Ltd. outside the submitted work. H. Goto reports personal fees from Daiichi Sankyo RD Novare Co. Ltd. during the conduct of the study and personal fees from Daiichi Sankyo RD Novare outside the submitted work. N. Wada reports personal fees from Daiichi Sankyo RD Novare Co. Ltd. during the conduct of the study and personal fees from Daiichi Sankyo RD Novare Co. Ltd. outside the submitted work. S. Tsutsumi reports other support from Daiichi Sankyo Co., Ltd. during the conduct of the study and other support from Daiichi Sankyo Co. Ltd. outside the submitted work. T. Ishikawa reports grants, personal fees, and nonfinancial support from Daiichi Sankyo Co., Ltd. during the conduct of the study; grants, personal fees, and nonfinancial support from Daiichi Sankyo Co., Ltd. outside the submitted work. S. Yamamoto reports personal fees from Daiichi Sankyo Co., Ltd. during the conduct of the study. No disclosures were reported by the other authors.

Y. Yoshihama: Conceptualization, data curation, formal analysis, investigation, methodology, writing—original draft, project administration, writing—review and editing. H. Namiki: Resources, investigation, compound development. T. Kato: Investigation. N. Shimazaki: Investigation. S. Takaishi: Investigation. K. Kadoshima-Yamaoka: Investigation, methodology. H. Yukinaga: Resources, investigation, methodology. N. Maeda: Resources, investigation, methodology. T. Shibutani: Resources, investigation, methodology. K. Fujimoto: Investigation, methodology, medicinal safety analysis. M. Hirasawa: Investigation. H. Goto: Investigation, methodology, RNA-seq analysis. N. Wada: Investigation, methodology, RNA-seq analysis. S. Tsutsumi: Conceptualization, supervision. Y. Hirota: Conceptualization, supervision. T. Ishikawa: Supervision. S. Yamamoto: Supervision.

The authors thank Drs. Eppenberger-Castori Serenella and Stefan Nicolet (University Hospital Basel) for providing TMA from patients with cancer.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

1.
Bryant
HE
,
Schultz
N
,
Thomas
HD
,
Parker
KM
,
Flower
D
,
Lopez
E
, et al
.
Specific killing of BRCA2-deficient tumours with inhibitors of poly(ADP-ribose) polymerase
.
Nature
2005
;
434
:
913
7
.
2.
Farmer
H
,
McCabe
N
,
Lord
CJ
,
Tutt
ANJ
,
Johnson
DA
,
Richardson
TB
, et al
.
Targeting the DNA repair defect in BRCA mutant cells as a therapeutic strategy
.
Nature
2005
;
434
:
917
21
.
3.
Muller
FL
,
Colla
S
,
Aquilanti
E
,
Manzo
VE
,
Genovese
G
,
Lee
J
, et al
.
Passenger deletions generate therapeutic vulnerabilities in cancer
.
Nature
2012
;
488
:
337
42
.
4.
Muller
FL
,
Aquilanti
EA
,
DePinho
RA
.
Collateral lethality: a new therapeutic strategy in oncology
.
Trends Cancer
2015
;
1
:
161
73
.
5.
Mavrakis
KJ
,
McDonald
ER
3rd
,
Schlabach
MR
,
Billy
E
,
Hoffman
GR
,
deWeck
A
, et al
.
Disordered methionine metabolism in MTAP/CDKN2A-deleted cancers leads to dependence on PRMT5
.
Science
2016
;
351
:
1208
13
.
6.
Kryukov
GV
,
Wilson
FH
,
Ruth
JR
,
Paulk
J
,
Tsherniak
A
,
Marlow
SE
, et al
.
MTAP deletion confers enhanced dependency on the PRMT5 arginine methyltransferase in cancer cells
.
Science
2016
;
351
:
1214
8
.
7.
Kohno
S
,
Linn
P
,
Nagatani
N
,
Watanabe
Y
,
Kumar
S
,
Soga
T
, et al
.
Pharmacologically targetable vulnerability in prostate cancer carrying RB1-SUCLA2 deletion
.
Oncogene
2020
;
39
:
5690
707
.
8.
Bergo
MO
,
Gavino
BJ
,
Steenbergen
R
,
Sturbois
B
,
Parlow
AF
,
Sanan
DA
, et al
.
Defining the importance of phosphatidylserine synthase 2 in mice
.
J Biol Chem
2002
;
277
:
47701
8
.
9.
Arikketh
D
,
Nelson
R
,
Vance
JE
.
Defining the importance of phosphatidylserine synthase-1 (PSS1): unexpected viability of PSS1-deficient mice
.
J Biol Chem
2008
;
283
:
12888
97
.
10.
Kimura
AK
,
Kimura
T
.
Phosphatidylserine biosynthesis pathways in lipid homeostasis: Toward resolution of the pending central issue for decades
.
FASEB J
2021
;
35
:
e21177
.
11.
Bepler
G
,
Gautam
A
,
McIntyre
LM
,
Beck
AF
,
Chervinsky
DS
,
Kim
YC
, et al
.
Prognostic significance of molecular genetic aberrations on chromosome segment 11p15.5 in non-small-cell lung cancer
.
J Clin Oncol
2002
;
20
:
1353
60
.
12.
Luan
M
,
Song
F
,
Qu
S
,
Meng
X
,
Ji
J
,
Duan
Y
, et al
.
Multi-omics integrative analysis and survival risk model construction of non-small cell lung cancer based on The Cancer Genome Atlas datasets
.
Oncol Lett
2020
;
20
:
58
.
13.
Wikman
H
,
Sielaff-Frimpong
B
,
Kropidlowski
J
,
Witzel
I
,
Milde-Langosch
K
,
Sauter
G
, et al
.
Clinical relevance of loss of 11p15 in primary and metastatic breast cancer: association with loss of PRKCDBP expression in brain metastases
.
PLoS One
2012
;
7
:
e47537
.
14.
Kozlowski
L
,
Filipowski
T
,
Rucinska
M
,
Pepinski
W
,
Janica
J
,
Skawronska
M
, et al
.
Loss of heterozygosity on chromosomes 2p, 3p, 18q21.3 and 11p15.5 as a poor prognostic factor in stage II and III (FIGO) cervical cancer treated by radiotherapy
.
Neoplasma
2006
;
53
:
440
3
.
15.
Liu
Y
,
Raheja
R
,
Yeh
N
,
Ciznadija
D
,
Pedraza
AM
,
Ozawa
T
, et al
.
TRIM3, a tumor suppressor linked to regulation of p21(Waf1/Cip1)
.
Oncogene
2014
;
33
:
308
15
.
16.
Moskaluk
CA
,
Rumpel
CA
.
Allelic deletion in 11p15 is a common occurrence in esophageal and gastric adenocarcinoma
.
Cancer
1998
;
83
:
232
9
.
17.
Mirlacher
M
,
Simon
R
.
Recipient block TMA technique
.
Methods Mol Biol
2010
;
664
:
37
44
.
18.
Tomohiro
S
,
Kawaguti
A
,
Kawabe
Y
,
Kitada
S
,
Kuge
O
.
Purification and characterization of human phosphatidylserine synthases 1 and 2
.
Biochem J
2009
;
418
:
421
9
.
19.
Namiki
H
,
Saitou
M
,
Matsui
S
,
Shibata
Y
,
Kawamoto
Y
,
Ichikawa
RIE
, et al
.;
Daiichi Sankyo Co., Ltd., assignee. Pyrrolopyrazole derivative
.
WO patent WO 2020/179859 A1. 5/3/2020
.
20.
Dobin
A
,
Davis
CA
,
Schlesinger
F
,
Drenkow
J
,
Zaleski
C
,
Jha
S
, et al
.
STAR: ultrafast universal RNA-seq aligner
.
Bioinformatics
2013
;
29
:
15
21
.
21.
Li
B
,
Ruotti
V
,
Stewart
RM
,
Thomson
JA
,
Dewey
CN
.
RNA-Seq gene expression estimation with read mapping uncertainty
.
Bioinformatics
2009
;
26
:
493
500
.
22.
Subramanian
A
,
Tamayo
P
,
Mootha
VK
,
Mukherjee
S
,
Ebert
BL
,
Gillette
MA
, et al
.
Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles
.
Proc Natl Acad Sci U S A
2005
;
102
:
15545
50
.
23.
Weiss
M
,
Blazek
K
,
Byrne
AJ
,
Perocheau
DP
,
Udalova
IA
.
IRF5 is a specific marker of inflammatory macrophages in vivo
.
Mediators Inflamm
2013
;
2013
:
245804
.
24.
Gopisetty
A
,
Bhattacharya
P
,
Haddad
C
,
Bruno
JC
,
Vasu
C
,
Miele
L
, et al
.
OX40L/Jagged1 cosignaling by GM-CSF–induced bone marrow-derived dendritic cells is required for the expansion of functional regulatory T cells
.
J Immunol
2013
;
190
:
5516
25
.
25.
Hamilton
JA
.
Colony stimulating factors and macrophage heterogeneity
.
Inflamm Regen
2011
;
31
:
228
36
.
26.
Liberzon
A
,
Birger
C
,
Thorvaldsdóttir
H
,
Ghandi
M
,
Mesirov
JP
,
Tamayo
P
.
The Molecular Signatures Database (MSigDB) hallmark gene set collection
.
Cell Syst
2015
;
1
:
417
25
.
27.
Radogna
F
,
Diederich
M
.
Stress-induced cellular responses in immunogenic cell death: implications for cancer immunotherapy
.
Biochem Pharmacol
2018
;
153
:
12
23
.
28.
Rufo
N
,
Garg
AD
,
Agostinis
P
.
The unfolded protein response in immunogenic cell death and cancer immunotherapy
.
Trends Cancer
2017
;
3
:
643
58
.
29.
Folger
O
,
Jerby
L
,
Frezza
C
,
Gottlieb
E
,
Ruppin
E
,
Shlomi
T
.
Predicting selective drug targets in cancer through metabolic networks
.
Mol Syst Biol
2011
;
7
:
501
.
30.
Hernández-Alvarez
MI
,
Sebastián
D
,
Vives
S
,
Ivanova
S
,
Bartoccioni
P
,
Kakimoto
P
, et al
.
Deficient endoplasmic reticulum-mitochondrial phosphatidylserine transfer causes liver disease
.
Cell
2019
;
177
:
881
95
.
31.
Patel
D
,
Witt
SN
.
Ethanolamine and phosphatidylethanolamine: partners in health and disease
.
Oxid Med Cell Longev
2017
;
2017
:
4829180
.
32.
Fu
S
,
Watkins
SM
,
Hotamisligil
GS
.
The role of endoplasmic reticulum in hepatic lipid homeostasis and stress signaling
.
Cell Metab
2012
;
15
:
623
34
.
33.
McGranahan
N
,
Swanton
C
.
Biological and therapeutic impact of intratumor heterogeneity in cancer evolution
.
Cancer Cell
2015
;
27
:
15
26
.

Supplementary data