An altered lipidome in tumors may affect not only tumor cells themselves but also their microenvironment. In this study, a lipidomics screen reveals increased amounts of phosphatidylserine (PS), particularly ether-PS (ePS), in murine mammary tumors compared with normal tissue. PS was produced by phosphatidylserine synthase 1 (PTDSS1), and depletion of Ptdss1 from tumor cells in mice reduced ePS levels accompanied by stunted tumor growth and decreased tumor-associated macrophage (TAM) abundance. Ptdss1-deficient tumor cells exposed less PS during apoptosis, which was recognized by the PS receptor MERTK. Mammary tumors in macrophage-specific Mertk−/− mice showed similarly suppressed growth and reduced TAM infiltration. Transcriptomic profiles of TAMs from Ptdss1-knockdown tumors and Mertk−/− TAMs revealed that macrophage proliferation was reduced when the Ptdss1/Mertk pathway was targeted. Moreover, PTDSS1 expression correlated positively with TAM abundance but negatively with breast carcinoma patient survival. PTDSS1 thus may be a target to modify tumor-promoting inflammation.

Significance:

This study shows that inhibiting the production of ether-phosphatidylserine by targeting phosphatidylserine synthase PTDSS1 limits tumor-associated macrophage expansion and breast tumor growth.

Cancer is driven by somatic mutations accumulating in cells that alter protein activity and consequently cellular signaling events. Considerable efforts have been undertaken to acquire genomic information of patients with cancer with the idea to identify signatures that inform clinical decisions towards precision cancer medicine (1). There are, however, caveats to this approach, because predicting a druggable mutation requires knowledge about the mutated protein and the intercellular and metabolic consequences targeting it would have (2). One such metabolic consequence is a modified lipidome. Tumors depend on increased lipid synthesis to meet their needs for energy and structural building blocks to enable replication (3). Moreover, lipophilic signaling molecules such as eicosanoids have been shown to promote various aspects of tumorigenesis (4). Tumor cells also frequently harbor alterations in membrane lipid composition. These alterations result in changes in membrane fluidity, which may promote invasion when reduced, or strengthen intra- and intercellular signaling when being locally increased, leading to the formation of membrane microdomains (5). Changes in plasma membrane lipid distribution affect drug uptake and the interaction of cancer cells with their cellular microenvironment (6). Concerning the latter, exposure of phosphatidylserine (PS) at the outer leaflet of the plasma membrane of melanoma cells was correlated with increased recognition by innate immune cells and pro-coagulant activity (7, 8). Changes in the tumor lipidome can also result in de novo generation or overproduction of lipid antigens that are recognized by natural killer T (NKT) cells, which, based on their phenotype, may suppress or support tumor growth (9).

Although a number of studies have characterized lipids as biomarkers in the plasma of cancer patients, there have been few systematic efforts to analyze the cancer lipidome and its impact on the tumor microenvironment. An elevated production of glucosylceramides was reported in breast cancer (10). Moreover, the mitochondrial lipidome in murine brain tumors was enriched in immature cardiolipin species (11), and ether phospholipid biosynthesis was increased in a number of human tumors (12). We set out to identify tumor-enriched lipids from autochthonous mouse mammary tumors compared with normal mammary glands, with the premise that altering their composition might interfere with tumor growth.

Animals

Female C57BL/6J mice expressing the polyoma virus middle T oncoprotein (PyMT) were used for primary tumor isolation compared with littermate controls. For orthotopic tumor transplantation studies, 1×106 murine mammary carcinoma cells derived from a PyMT tumor were injected into two mammary glands of female C57BL/6J WT mice, B6.129S6-Del(3Cd1d2-Cd1d1)1Sbp/J (CD1d KO) mice (The Jackson Laboratory), Csf1r-CreMertkf/f, or Csf1r-Cre+Mertkf/f mice (13), respectively. The impact of cytotoxic immune cells on tumor growth was tested by intraperitoneal injection with anti-CD8 (YTS 169.4, BioXCell), anti-NK1.1 (PK136, BioXCell) neutralizing antibodies or respective isotype controls (BioXCell) diluted in sterile 0.9% NaCl solution twice weekly starting on week 2 after tumor cell injection. Tumor growth was monitored for 3, 5, or 10 weeks by caliper measurements. At the respective endpoint, mice were euthanized and cardially perfused, and tumors were harvested for downstream analyses.

Lipidomics screening tumor-enriched lipids

Tumors were harvested from 18 weeks old PyMT mice, and normal breast tissue was harvested from littermate controls of the same age. Tissues were snap-frozen with liquid nitrogen and stored at −80°C until further use. Tissue were ground under liquid nitrogen, resolved in PBS and sonicated. Lipid isolation from tissue homogenates reflecting 10 mg original tissue was done using a modified-Folch method (organic solvent consisting of 83% chloroform, 15% methanol, and 2% HCl). The organic phase containing the lipids was dried under a stream of nitrogen. 200 µL streptavidin-coupled magnetic beads (Invitrogen) of a 10 mg/mL stock solution were washed with PBS and resuspended in 600 µL 0.5% PBS/BSA. Lipids were enriched by CD1d binding essentially as described previously (14). Afterwards, lipids were eluted, dried under liquid nitrogen, and stored at −80°C until use. The nontargeted lipidomics screen was performed via LC-QTOF/MS at Metabolomics Discovery GmbH (now Metabolon Inc.). Metabolites in the range of 50 to 1,700 Da with an accuracy up to 1 to 2 ppm and a resolution of mass/Δmass = 40.000 were analyzed, annotated according to their accurate mass and subsequent sum formula prediction. Metabolites that were not annotated in the LC/MS analyses were listed according to their accurate mass and retention time (accurate mass@retention time, e.g., 230.9478@3.1). All features not robustly measured throughout the experiment were filtered out. Criteria were higher concentration in blank samples than in biological samples, presence in 80% of all samples, and presence in 33% of one experimental group. 160 differentially expressed metabolites (local P ≤ 0.05) were identified. 94 metabolites were identified as being upregulated based on a fold change of >1.5, followed by correcting for multiple testing (FDR).

Lipid quantification by mass spectrometry

Targeted phospholipid analysis was performed as follows. Samples were analyzed by ESI/MS-MS in positive ion mode by direct flow injection using the analytical setup and data analysis algorithms essentially as described previously (15). A precursor ion scan of m/z 184 specific for phosphocholine containing lipids was used for phosphatidylcholine (PC) and lysophosphatidylcholine (LPC). Fragment ions of m/z 364, 390, and 392 were used for PE P-16:0, PE P-18:1, and PE P-18:0 species, respectively. Neutral loss scan of m/z 185 was used for PS.

PS ether lipids were identified as follows. For targeted screening the [M–H]-87 fragment ions of the PS species were used. Lysis was done by adding 10 µL/mg 1:1 (MeOH:ButOH) with 5 mmol/L ammonium formate. Next, samples were sonicated for 1 hour in a water bath and centrifuged for 15 minutes at 15,000 × g. Forty microliters of sample were mixed with 10 µL SPLASH LIPIDOMIX Mass Spec Standard (Avanti Polar Lipids). LC/MS-MS was performed on an Agilent 1290 Infinity LC system (Agilent) coupled to a QTrap5500 mass spectrometer (Sciex). Ion source parameters were as follows: CUR 35 psi, CAD medium, ion spray voltage: 4,500 V, TEM: 300°C, GS1: 30 psi, GS2: 40 psi. Reversed-phase LC separation was performed using a Kinetex C18 column (2.1 × 100 mm, 1.7 µm; Phenomenex). Gradient elution was performed with isopropanol/methanol/water (5:1:4) containing 0.2% formic acid, 0.1% InfinityLab Deactivator Additive (Agilent Technologies), 0.028% ammonia (mobile phase A), and 90% isopropanol containing the three additives indicated above (mobile phase B). Compounds were eluted with a flow rate of 0.15 mL/min and with the following 25 minutes gradient: 0 minute 30% B, 0 to 6 minutes 50% B, 6 to 17 minutes 80% B, 17 to 17.5 minutes 95% B, 17.5 to 20 minutes 95% B, 20 to 20.5 minutes 5% B, 20.5 to 21 minutes 5% B, 21 to 21.5 minutes 30% B, and 21.5 to 25 minutes 30% B. Analyst 1.6.2 and MultiQuant 3.0 (both from Sciex) were used for data acquisition and analysis, respectively. Quantification of PS(O-18:0/0:0) was performed against a calibration curve with authentic standard.

Synthesis of PS(O-18:0/0:0)/ST-1943

The convergent synthesis of PS(O-18:0/0:0)/ST-1943 was accomplished from two different building blocks. One is the orthogonally protected phosphoamidite derivative of l-serine (16, 17). The second one is the tert.-butyldimethylsilyl-protected primary alcohol (4) synthesized in four steps according to reference methods (18, 19). After coupling of those two building blocks (2) and (4) (ref. 17) and following full deprotection under acidic conditions resulted in the phospholipid serine probe ST-1943. Details of the synthesis are available in supporting information (Supplementary Fig. S2).

Analysis of publicly available human data sets

The Cancer Genome Atlas (TCGA) colon adenocarcinoma data set (20), the METABRIC data set (21), and a data set of patients with invasive mammary carcinoma (22) were downloaded from cBioPortal for Cancer Genomics (http://www.cbioportal.org) including associated clinical data. Gene expression values and clinical data were statistically evaluated using GraphPad Prism v9.

IHC of human and murine tumor tissue

Analysis of macrophage phenotypes in murine transplanted PyMT tumors, and PTDSS1 expression and macrophage infiltrates in invasive breast cancer tissue microarrays specimens provided by the Cooperative Human Tissue Network and the Cancer Diagnosis Program (other investigators may have received specimens from the same subjects) was done by PhenOptics as described previously (23). Opal Fluorescent IHC kits (Akoya Biosciences) were used according to the manufacturer's instructions. Murine PFA-fixed tumor tissue sections were stained with primary antibodies targeting pan-Cytokeratin (Abcam, ab27988), CD163 (Abcam, ab182422), CD206 (Abcam, ab64693), ALOX15 (Abcam, ab244205), TNFα (Abcam, ab6671), IL-12p70 (Thermo Fisher Scientific, MM120), vimentin (Abcam, ab92547), and Ki67 (Abcam, ab16667). Human tissue microarrays prepared paraffin sections were stained with primary antibodies against PTDSS1 (Sigma, HPA016852), pan-Cytokeratin (Abcam, ab7753), and CD163 (Abcam, ab182422). Nuclei were counterstained with DAPI. The BOND RX Automated IHC Research Stainer (Leica Biosystems) was used for staining. Samples were acquired at 20× magnification using the Vectra3 automated quantitative pathology imaging system (Akoya Biosciences). InForm v2.4 (Akoya Biosciences) was used to quantify the percentage of CD163+ macrophages and macrophage subsets in murine tumor stroma after tissue segmentation, or the percentage of CD163+ macrophage and PTDSS1-expressing tumor cells in human tissue microarrays. Human cancer cores were evaluated based on tissue integrity and quality after staining. On the basis of these criteria, 169 individual cores were suitable for analysis.

Cell culture

Murine primary PyMT mammary carcinoma cells (PyMT) were cultured in DMEM, murine IC-21 macrophages (ATCC Catalog No. TIB-186, RRID:CVCL_3726) and Jurkat cells were cultured in RPMI1640, each supplemented with 10% FCS (Capricorn Scientific), 100 U/mL penicillin, and 100 µg/mL streptomycin (PAA laboratories). For IC-21 cells, 1% sodium pyruvate, 1% nonessential amino acids, and 10 mmol/L HEPES (Sigma Aldrich) were additionally added. Cells were kept at 37°C in a humidified atmosphere of 95% air /5% CO2 with medium exchanges every 2 to 3 days. IC-21 cells were used within 1 year after purchase and never expanded beyond passage 9. Authentication for Jurkat cells was not done. All cell lines were routinely tested for Mycoplasma contamination using Venor GeM Mycoplasma Detection Kit (Sigma) at least bimonthly when in culture.

Lentiviral Ptdss1-knockdown

pGFP-C-shLenti plasmids (OriGene) were used to knockdown Ptdss1 expression in PyMT cells. HEK293T cells were transfected using jetPRIME reagents (Polyplus transfections), with four different plasmids encoding shRNAs against Ptdss1 and a nontargeting shRNA control, as well as vectors encoding packaging and envelope constructs for lentiviral generation. The virus-containing supernatants were harvested, stored at −80°C, and used to infect PyMT mammary carcinoma cells. Puromycin (2.5 µg/mL) was used to enrich positive cells. Clonal selection was not performed to keep respective artifacts at a minimum. Two different Ptdss1 knockdown cell lines were used for all experiments to further exclude off-target artifacts.

RNA extraction and qRT-PCR

RNA from tumor samples were isolated using the PeqGold RNAPure protocol (Peqlab Biotechnologie) and transcribed into cDNA using the Maxima First Strand cDNA Synthesis Kit (Thermo Fisher Scientific). qRT-PCR was performed using PowerUp SYBR Green Master Mix and the QuantStudio 5 Realtime-PCR system (Thermo Fisher Scientific). Quantitect primer assays (Qiagen) were used to detect murine Far1, Far2, Agps, Gnpat, Ptdss1, Ptdss2, Il1b, Il6, Il10, Il23a, Arg1, Chil3, Tgm2, Tgfb1, Ki67, Ccni, Pole, Adgre5, and Timp1. Other murine primer sequences (obtained from biomers.net) were: Rps27a F: 5′-GAC CCT TAC GGG GAA AAC CAT-3′, R: 5′-AGA CAA AGT CCG GCC ATC TTC-3′; Mertk F: 5´-TGC GTT TAA TCA CAC CAT TGG A-3´, R: 5´-TGC CCC GAG CAA TTC CTT TC-3´; Stabilin2 F: 5´-GTT GCT TGT GCA AAA TGC CTG-3´, R: 5´-GCA CTC CGT CTT GAT GGT TAG AG-3´, Tim4 F:5´-CCG GTG ACT TTG CCT TGT CAT-3´, R: 5´-CTC TGC ATT GCA CTT GGA ATT G-3´, Bai1 F: 5´-TTG CTC CAC TCC TGC TGT TAC-3´, R: 5´- GTA GCC GAA CTT TCC CTG-3´, Tnfa F: 5´-TGC CTA TGT CTC AGC CTC TT-3´, R: 5´-GAG GCC ATT TGG GAA CTT CT-3´, Tyro3 F: 5´-GCC TCC AAA TTG CCC GTC A A-3´, R: 5´-CCA GCA CTG GTA CAT GAG ATC -3´, Axl F: 5´-CGG TAG TAA TCC CCG TTG TAG A-3´, R: 5´-ATG GCC GAC ATT GCC AGT G-3´. Relative mRNA expression was calculated using the ΔΔCt method and normalized to respective control RNAs indicated.

PTDSS1 activity assay

PTDSS1 activity was determined in endoplasmic reticulum (ER) fractions of control versus Ptdss1 knockdown PyMT cells enriched using Minute ER Enrichment Kits (MoBiTec). A published PTDSS1 activity assay with modifications was performed (24). Enriched ER fractions were prepared in an assay buffer containing 50 mmol/L Hepes/NaOH (pH 7.5), 5 mmol/L CaCl2, and 1 mmol/L L-serine. Different amounts of ER fractions were added to assay buffer containing 10 mg/mL NBD-PE (18:1–6:1 NBP PE; from Avanti Polar Lipids). Altered NBD fluorescence over time, indicating headgroup exchange-dependent lipid layer structural changes, were recorded at 37°C using an Infinite 200 PRO plate reader (Tecan).

Proliferation assay

The IncuCyte R S3 live-cell analysis system (Sartorius) was used to study proliferation of control and Ptdss1-knockdown PyMT cells. Images were taken every 4 hours and the doubling time of cells was calculated as follows: doubling time = [duration + log2)/(log(final concentration) – log(initial concentration)].

Spheroid generation

Spheroids generated from PyMT cells were initiated using 1 × 104 cells/mL and were cultured in cell repellent 96-well plates (Greiner Bio-One GmbH) with a final concentration of 1.5% growth factor reduced Matrigel (Corning Costar) for 10 days. Spheroid size was acquired with a Carl Zeiss Axiovert microscope and diameters were determined using AxioVision 40 software (Carl Zeiss AG).

Lipid-loaded CD1d tetramer generation

Mouse CD1d-biotin monomers were obtained from the NIH Tetramer Core Facility. Tetramers were generated with fluorochrome-labeled streptavidin (BD Biosciences) using standard procedures as suggested by the NIH Tetramer Core Facility. Briefly, 2 µg lyso-ether-PS (ePS)/ST-1943 were dried under nitrogen, resolved in 45 µL TBS, pH 6, for 2 minutes at 80°C in an ultrasound bath. Afterwards, 3 µL saposin B (1 mg/mL stock solution; a kind gift from Robert P. Doyle, Syracuse University, Syracuse, NY; ref. 25) was added, followed by incubation in an ultrasound bath for 10 minutes at room temperature. To this mixture, 3 µL CD1d/D-biotin (2 mg/mL stock solution) were added mixed and incubated for 2 hours at 37°C in the dark with gentle shaking. During this time 0.95 µL streptavidin-PE-CF594 was added every 10 minutes for a total of 10 times. The resulting lipid-loaded CD1d tetramers were stored at 4°C until use.

Flow cytometry

Single cell suspensions of tumors were generated using the Mouse Tumor Dissociation Kit and the gentleMACS dissociator (Miltenyi Biotec). This method, when applied to transplanted tumors, efficiently isolates live immune cells, while being less efficient in maintaining live tumor cells. All single cell suspensions were stained with fluorochrome-coupled antibodies and analyzed by flow cytometry using an LSRII Fortessa cell analyzer or sorted using a FACS Aria III cell sorter (both from BD Biosciences). Data were analyzed using FlowJo software VX (TreeStar). Antibodies were titrated to determine optimal concentrations. For single-color compensation CompBeads (BD Bioscience) were used to create multicolor compensation matrices. Flow Cytometry Absolute Count Standard (BangsLabs) beads were used as an internal counting standard. Cells were blocked with 2% Fc Receptor Binding Inhibitor (Miltenyi Biotec) in PBS for 10 minutes on ice. Afterwards, cells were stained to analyze immune cell composition. The following antibodies were used to characterize mouse tumors. anti-CD3-PE-CF594 (145–2C11), anti-CD4-BV711 (GK1.5), anti-CD8-BV650 (53–6.7), anti-CD11c-Alexa Fluor 700 (HL3), anti-CD326-BV711 (G8.8), anti-Ly6C-PerCP-Cy5.5 (AL-21), anti-NK1.1-BV510 (PK136; BD Biosciences), anti-CD31-PE-Cy7 (390, eBioscience), anti-CD45-VioBlue (30-F11.1), anti-CD90.2-PE (30–H12), anti-MHC-II-APC (M5/114.15.2; Miltenyi Biotec), anti-CD11b-BV605 (M1/70), anti-F4/80-PE-Cy7 (BM8), anti-GITR-FITC (DTA-1), and anti-Ly6G-APC-Cy7 (1A8; BioLegend). 7-AAD was used for dead cell exclusion. Macrophages were FACS-sorted from mouse tumors as CD45+, CD11b+, Ly6C, Ly6G, F4/80+ cells. To detect NKT cells in mice, anti-CD45-Alexa Fluor 700 (30–F11), anti-CD3-PE-Cy5 (145–2C11), anti-NK1.1-BV510 (PK136; BD Biosciences), and lyso-ePS-loaded CD1d tetramer-PE-CF594 were used. To stain for intracellular mediators, CD3-PE-CF594 (145–2C11), anti-CD4-BV711 (GK1.5), anti-CD8-BV650 (53–6.7), anti-CD45-AlexaFluor700 (30–F11), anti-CD326-BV711 (G8.8), anti-IFNγ-BV421 (XMG1.2), anti-NK1.1-BV510 (PK136; BD Biosciences), anti-CD11b-BV605 (M1/70), anti-F4/80-PE-Cy7 (BM8), and anti-Perforin-APC (S16009B; BioLegend) were used. Cells were stained with surface markers, and fixed with Cytofix/Cytoperm buffer (BD Biosciences) for a maximum of 10 minutes on ice, followed by washing and permeabilization with Perm/Wash buffer (BD Biosciences). Permeabilized cells were blocked with Fc Receptor Binding Inhibitor for 15 minutes prior to intracellular staining. Cell death and PS exposure on PyMT cells or IC-21 macrophages was analyzed using the Annexin V/propidium iodide (PI; Sigma) method. Annexin V-APC (BD Biosciences) and PI were added to cells in Annexin V buffer for 15 minutes on ice before acquisition. For FACS-FRET analysis, Jurkat cells were incubated with 2 µg/mL NBD-labeled PS (Avanti Polar Lipids) with or without addition of lyso-PS or lyso-ePS at 1:1 or 1:5 ratios for 2 hours. Afterwards, apoptosis was induced with 0.5 µg/mL staurosporin (Sts; Sigma) for 3 hours. Cells were then either left unstained or stained with Annexin V-V450 or Annexin V-BV421 (both from BioLegend). 405 and 488 nm lasers were used for excitation in the FRET assay, NBD fluorescence, and the FRET signal were detected using 525/50 nm filters, and V450 or BV421 were detected using 450/50 nm filters. FRET efficiency was calculated based on fluorescence medians of the above-mentioned fluorochromes using the FRET calculator protocol provided by Ujlaky-Nagy and colleagues (26). Single-positive as well as fluorescence negative cells served as background controls.

Cytokine analysis

Mouse TNFα protein concentrations in cell supernatants were measured by FACS using Cytometric Bead Array Flex Sets (BD Biosciences). Samples were acquired with the LSR Fortessa flow cytometer (BD Biosciences) and analyzed with BD Bioscience's FCAP software (V3).

Induction of cell death and generation of conditioned media

PyMT cells or Jurkat cells were treated with 0.5 µg/mL staurosporin for the times indicated to induce cell death. Apoptotic cell-conditioned (ACM) and viable cell-conditioned media (VCM) were generated from control and Ptdss1-knockdown PyMT cells at 90% confluency. Cells were treated with 0.5 µg/mL Sts for 1 hour, followed by washing with PBS for three times to remove Sts, and addition of fresh media for 24 hours to generate ACM. To generate VCM cells were cultivated in media for 24 hours. Afterwards, supernatants were harvested and cells, but not apoptotic bodies were removed by centrifugation. Conditioned media were stored at −80°C until use.

IC-21 macrophage stimulation

IC-21 macrophages were plated in 6-well plates at a confluency of 80% and were subsequently stimulated with ACM or VCM, generated as described above for the time points indicated. In some experiments, IC-21 cells were pre-incubated with 50 nmol/L of the MERTK inhibitor UNC2025. To reconstitute PS levels, conditioned media were preincubated with 1 µg/mL lyso-PS or lyso-ePS for 10 minutes in an ultrasound bath before adding them to cells. Cells were harvested for mRNA analysis and supernatants were stored at −80°C until use for cytokine analysis or addition to PyMT cells. In proliferation experiments, IC-21 macrophages were prelabelled with cell proliferation dye eFluor 670 (eBioscience) according to the manufacturer's instructions. Proliferation was determined by dye dilution and cell counting, both via flow cytometry. To analyze PyMT cell phagocytosis, PyMT cells were labeled with PKH67 and IC-21 macrophages were labeled with PKH26 (both from Sigma). Cell death in PyMT cells was induced with 0.5 µg/mL Sts for 24 hours, and PyMT were incubated with IC-21 macrophages for 1 hour before analysis of phagocytosing (PKH67+ PKH26+) IC-21 macrophages via flow cytometry.

RNA sequencing of tumor-associated macrophages

mRNA-isolation and cDNA transcription were performed immediately after tumor-associated macrophage (TAM) sorting using SMART-Seq v4 Ultra Low Input RNA Kit for Sequencing (Takara) according to the manufacturer's instructions. cDNA was amplified by 16 PCR cycles and samples were purified after cDNA and library synthesis with AMPure XP beads for PCR Purification (Beckman Coulter). For the determination of cDNA content, a Qubit 3.0 Fluorometer in combination with Qubit 1× dsDNA HS Assay Kit (both Thermo Fisher Scientific) were used. Fragment sizes were determined using High Sensitivity DNA Kit with a 2100 Bioanalyzer (both Agilent). The Covaris M220 system was used for controlled cDNA shearing (2 minutes at 20°C, Peak Power 75W, 20% Duty Factor, Burst Cycle set at 200). Library preparation was performed with SMARTer ThruPLEX DNA-Seq Kit (Takara). Ten samples were subjected to an indexed single-end sequencing run with 75 cycles on an Illumina NextSeq 500 system. NGS data analysis was performed using the SeqBox ecosystem. Briefly, after adapter trimming with skewer, reads were mapped to the murine reference genome (mm10) using STAR. Gene level quantification by RSEM preceded the differential expression analysis by DESeq2.

Heatmap generation

MORPHEUS matrix visualization and analysis software (https://software.broadinstitute.org/morpheus/) was used for heatmap generation. One minus Pearson's correlation was used for K means clustering of rows.

Gene set enrichment analysis

Gene set enrichment analysis (GSEA) was performed with a shared gene signature between macrophages from Ptdss1-KD tumors and Mertk-deficient TAMs using the GSEA module (27) on the GenePattern platform (28).

Statistical analysis

Unless stated otherwise, data are presented as means ± SEM. Statistical comparisons between two groups were performed using Mann–Whitney test, paired or unpaired two-tailed Student t test as indicated. One- or two-way ANOVA, or multiple t tests, followed by appropriate posttests was used for multiple comparisons. Data were pre-analyzed to determine normal distribution and equal variance with D'Agostino–Pearson omnibus normality test. Normalized data were analyzed using one-sample t test. Statistical survival analysis was performed using the log-rank test. Correlation was analyzed using Spearman test for comparing continuous, or using point-biserial correlation for comparing continuous and binary categorical variables. Statistical analysis was performed with GraphPad Prism v9, or R v3.6.1 for point-biserial correlation. Differences were considered significant if *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001. No statistical test was used to predetermine sample size, and all samples were included in the analysis. Details on statistical tests used in individual experiments can be found in the figure legends.

Study approval

Mouse care and experiments involving mice were approved by and conducted according to the guidelines of the Hessian animal care and use committee (FU/1020, FU/1122, FU/1181, FU/2014, F123/1049).

Data availability statement

The transcriptomic datasets generated during and/or analyzed during this study are available at GEO: GSE157241. All other datasets generated during and/or analyzed during this study are available from the corresponding author on reasonable request.

Enrichment of ePS in murine mammary tumors

To test for tumor-enriched lipids, we harvested untransformed mammary glands and polyoma virus middle T oncogene (PyMT) mammary tumors, and measured lipids via high-resolution LC-QTOF/MS analysis (Fig. 1A and B; Supplementary Table S1; Supplementary Fig. S1). Fifteen significantly upregulated metabolites in tumors compared with the mammary gland could be annotated (Fig. 1C). Among those lipids, phospholipids were enriched, particularly PS species. Moreover, a large proportion of enriched PS species were ether lipids, both alkyl (O) and alkenyl (P, plasmalogens) species (Fig. 1C). Quantitative lipid analysis of mammary glands and PyMT mammary tumors by MS revealed increased abundance of total ether lipids and total PS species (Fig. 1D and E). Importantly, PS levels were significantly elevated relative to other phospholipids, indicating a specific increase in PS (Fig. 1F). There was no particular enrichment of short versus long chain or saturated versus unsaturated PS species (Fig. 1G). In turn, semiquantitative lipidomics suggested a particular enrichment of lyso ePS species, including PS(O-18:0/0:0) identified in the lipidomics screen (Fig. 1C), as well as ePS species containing saturated or monounsaturated fatty acids (Fig. 1H). To evaluate changes in ePS species quantitatively, PS(O-18:0/0:0) was synthesized (ST-1943; Supplementary Fig. S2) and used as a standard for LC/MS-MS, confirming the induction of PS(O-18:0/0:0) in mammary tumors compared with mammary glands (Fig. 1I). These data suggest that PyMT mammary tumors contain higher levels of PS, particularly ePS species.

Figure 1.

Abundance of ether-linked PS species in PyMT mammary tumors. A, Schematic diagram of lipidomics screen. Lipids isolated from PyMT murine mammary tumors or normal breast tissue (n = 6 each) were subjected to mass spectrometric analysis. B, Volcano plot shows abundance of lipids in tumors versus mammary glands. Dotted gray lines separate differentially expressed lipids (local P ≤ 0.05; log2-fold expression ± 0.5). C, Heatmap with names and relative expression of lipids significantly upregulated (fold change 1.5) in tumors versus mammary glands (q ≤ 0.05). Red, PS species. Ether lipids are indicated by P (alkenyl) or O (alkyl). D–I, Bulk lipids isolated from mouse mammary glands or mammary tumors were isolated and subjected to mass spectrometry analysis. Total ether lipid content (D), total PS content (E), PS content relative to other phospholipids (F), expression levels of individual PS species (G), or ether-PS species (H), as well as PS(O-18:0/0:0; I) are shown. Data are means ± SEM; mammary gland n = 9, mammary tumor n = 6 (D, E, and G); mammary gland n = 7, mammary tumor n = 6 (F); n = 4 each (H and I). D–F and I, Mann–Whitney test; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 1.

Abundance of ether-linked PS species in PyMT mammary tumors. A, Schematic diagram of lipidomics screen. Lipids isolated from PyMT murine mammary tumors or normal breast tissue (n = 6 each) were subjected to mass spectrometric analysis. B, Volcano plot shows abundance of lipids in tumors versus mammary glands. Dotted gray lines separate differentially expressed lipids (local P ≤ 0.05; log2-fold expression ± 0.5). C, Heatmap with names and relative expression of lipids significantly upregulated (fold change 1.5) in tumors versus mammary glands (q ≤ 0.05). Red, PS species. Ether lipids are indicated by P (alkenyl) or O (alkyl). D–I, Bulk lipids isolated from mouse mammary glands or mammary tumors were isolated and subjected to mass spectrometry analysis. Total ether lipid content (D), total PS content (E), PS content relative to other phospholipids (F), expression levels of individual PS species (G), or ether-PS species (H), as well as PS(O-18:0/0:0; I) are shown. Data are means ± SEM; mammary gland n = 9, mammary tumor n = 6 (D, E, and G); mammary gland n = 7, mammary tumor n = 6 (F); n = 4 each (H and I). D–F and I, Mann–Whitney test; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

PTDSS1-knockdown reduces ePS levels

In mammals, PS is generated by two PS synthases (PTDSS), PTDSS1 and PTDSS2, which catalyze a base-exchange reaction in the ER, where the head group of PC and PE is replaced by serine. PS degradation in mitochondria occurs by decarboxylating serine to ethanolamine by PS decarboxylase (PISD), generating PE (Fig. 2A). Although data on ePS synthesis are not available, an involvement of PTDSS isoenzymes appeared likely. In silico analysis using the COSMIC database revealed a frequent overexpression (26.72%) and copy-number gain (10.79%) of PTDSS1 in patients with breast cancer (29), which was not observed for PTDSS2 and PISD. This expression pattern was reproduced in mouse mammary tissue compared with PyMT tumors (Fig. 2B). We therefore asked whether reducing PTDSS1 levels in PyMT cells would affect ePS levels and tumor growth. We used an in-house generated PyMT tumor-derived cell line and produced stable Ptdss1 knockdown (Ptdss1-KD) PyMT cells by lentiviral shRNA transduction. Successful Ptdss1-knockdown was confirmed in four lines (Fig. 2C). One line (line A) showing compensatory upregulation of Ptdss2 (Supplementary Fig. S3A), and another (line C) demonstrating unspecific downregulation of ether lipid metabolizing enzymes (Supplementary Fig. S3B) were discontinued. Cell lines B and D displayed reduced PTDSS1 activity compared with a control line transduced with nontargeting shRNA (Fig. 2D). Moreover, Ptdss1-deficient PyMT cell lines showed reduced ePS abundance (Fig. 2E; Supplementary Fig. S3C), whereas there were no significant differences in PS species (Fig. 2F). Ptdss1-KD PyMT cells neither showed reduced growth rates in 2D or 3D cultures (Fig. 2G and H), nor altered viability (Fig. 2I) in vitro. Rather, one Ptdss1-KD PyMT cell line showed an increased growth rate (Fig. 2G and H), which was not reproducibly observed upon Ptdss1-KD (Supplementary Fig. S3D). Thus, Ptdss1-knockdown in PyMT cells reduced ePS levels, without suppressing tumor cell proliferation or death.

Figure 2.

Ptdss1-KD reduces ePS levels. A, Schematic depiction of PS metabolism in humans. B, Expression of PTDSS1, PTDSS2, and PISD in mouse mammary glands compared with mammary tumors. CF, PyMT cells were lentivirally transduced with nontargeting shRNA (Sh-Ctrl/Ctrl) or four different shRNAs targeting Ptdss1 (Sh-Ptdss1A-D/Ptdss1-KD). C–F, Expression of Ptdss1 at mRNA level (C), PTDSS1 activity (D), ePS (E) or PS (F) levels are shown. G, Proliferation of Ctrl or Ptdss1-KD PyMT cells (n = 3). H and I, Ctrl or Ptdss1-KD PyMT cells were grown as 3D spheroids for 10 days. Representative spheroid images (scale bar, 200 µm), spheroid diameter (10 spheroids of three experiments each; H) and cell viability determined by Annexin V/PI (I) are shown. Data are means ± SEM. B, Multiple t tests with FDR correction. C, D, and H, One-way ANOVA with FDR correction. E and F, Mann–Whitney test. *, P < 0.05; ***, P < 0.001.

Figure 2.

Ptdss1-KD reduces ePS levels. A, Schematic depiction of PS metabolism in humans. B, Expression of PTDSS1, PTDSS2, and PISD in mouse mammary glands compared with mammary tumors. CF, PyMT cells were lentivirally transduced with nontargeting shRNA (Sh-Ctrl/Ctrl) or four different shRNAs targeting Ptdss1 (Sh-Ptdss1A-D/Ptdss1-KD). C–F, Expression of Ptdss1 at mRNA level (C), PTDSS1 activity (D), ePS (E) or PS (F) levels are shown. G, Proliferation of Ctrl or Ptdss1-KD PyMT cells (n = 3). H and I, Ctrl or Ptdss1-KD PyMT cells were grown as 3D spheroids for 10 days. Representative spheroid images (scale bar, 200 µm), spheroid diameter (10 spheroids of three experiments each; H) and cell viability determined by Annexin V/PI (I) are shown. Data are means ± SEM. B, Multiple t tests with FDR correction. C, D, and H, One-way ANOVA with FDR correction. E and F, Mann–Whitney test. *, P < 0.05; ***, P < 0.001.

Close modal

Ptdss1-KD attenuates tumor growth and reduces TAM abundance

To analyze if knocking down Ptdss1 affects tumor growth in vivo, Ptdss1-deficient cells compared with control PyMT cells were orthotopically grafted into the mammary fat pad of syngenic mice and monitored for 5 weeks (Fig. 3A). Although initially growing at a similar rate than control tumors, growth of Ptdss1-deficient tumors was hampered from approximately week 3 onwards (Fig. 3B). As a result, tumors of both Ptdss1-KD lines were smaller than control tumors after 5 weeks (Fig. 3C and D). When tumors were monitored for another 5 weeks, Ptdss1-KD tumors maintained lower growth rates than control tumors (Fig. 3E). This resulted in improved survival of mice, with nearly all mice receiving control PyMT cells having to be sacrificed due to reaching ethical endpoints, whereas all mice with Ptdss1-deficient tumors survived (Fig. 3F). Reduction of Ptdss1-KD tumor growth 3 weeks after initiation suggested an involvement of the immune system. We first investigated a role of NKT cells, which can promote tumor growth in response to lipid antigens (9). Using lipid-loaded fluorochrome-coupled CD1d tetramers in flow cytometry indicated the presence of lyso-ePS-recognizing NKT cells in control tumors in mice, which were reduced in Ptdss1-KD PyMT tumors (Fig. 3G and H). However, lyso-ePS-recognizing NKT cells appeared not to be involved in PTDSS1-dependent tumor growth as elucidated by implanting control and Ptdss1-KD PyMT cells in Cd1d-deficient mice (Fig. 3I), despite Ptdss1-dependent production of so far undescribed NKT lipid antigens in tumors. To assess other immune mechanisms that account for the reduction in tumor growth following Ptdss1 KD, we analyzed tumor immune composition by flow cytometry, 5 weeks after tumor initiation (Supplementary Fig. S4A). The most consistent change in the tumor immune infiltrate in Ptdss1-KD tumors was a reduction in TAMs, which was observed using two Ptdss1-deficient PyMT cell lines and CD1d-deficient mice (Fig. 3JL). A nonconsistent increase in CD8+ T cells and NK cells was observed (Fig. 3JL). However, these cells showed enhanced activation determined by analyzing IFNγ expression, whereas CD4+ T cells did not (Fig. 3M). Interestingly, depleting CD8+ T cells or NK cells with neutralizing antibodies (Supplementary Fig. S4B) did not revert the growth of Ptdss1-KD tumors to the control level (Fig. 3N). Thus, PTDSS1 expression was required to sustain tumor growth in mice, which was associated with altered TAM infiltration.

Figure 3.

Ptdss1 deficiency impairs tumor growth and reduces TAMs. PyMT cells lentivirally transduced with nontargeting shRNA (Ctrl) or shRNA-targeting Ptdss1 (Ptdss1-KD) were engrafted into syngenic WT (B–G and J–N) or CD1d-deficient (H and I) mice, and tumors were maintained for 5 or 10 weeks. A–D, Experimental outline (A), tumor diameter over time (B), and tumor diameters at 5-week endpoint (C and D) are shown. E and F, Tumor diameter over time (E) and animal survival (ethical endpoint; F) are shown. G and H, Representative FACS plots (G) and the quantification (H) show abundance of lyso-ePS reactive NKT cells. I, Tumor diameter over time in CD1d-deficient mice is shown. J–L, Tumor-infiltrating immune cells within total living cells determined by flow cytometry. M, Intracellular perforin (Prf1) and IFNγ expression by CD4+ T cells, CD8+ T cells, and NK cells determined by flow cytometry. N, Tumor diameter over time after twice-weekly injection of anti-CD8, anti-NK1.1 neutralizing antibodies, or isotype control is displayed. One experiment out of two with 5 mice per group is shown. Data are means ± SEM. C, D, and H, Mann–Whitney test. I–M, Multiple t tests with FDR correction. F, Log-rank test. *, P < 0.05; **, P < 0.01.

Figure 3.

Ptdss1 deficiency impairs tumor growth and reduces TAMs. PyMT cells lentivirally transduced with nontargeting shRNA (Ctrl) or shRNA-targeting Ptdss1 (Ptdss1-KD) were engrafted into syngenic WT (B–G and J–N) or CD1d-deficient (H and I) mice, and tumors were maintained for 5 or 10 weeks. A–D, Experimental outline (A), tumor diameter over time (B), and tumor diameters at 5-week endpoint (C and D) are shown. E and F, Tumor diameter over time (E) and animal survival (ethical endpoint; F) are shown. G and H, Representative FACS plots (G) and the quantification (H) show abundance of lyso-ePS reactive NKT cells. I, Tumor diameter over time in CD1d-deficient mice is shown. J–L, Tumor-infiltrating immune cells within total living cells determined by flow cytometry. M, Intracellular perforin (Prf1) and IFNγ expression by CD4+ T cells, CD8+ T cells, and NK cells determined by flow cytometry. N, Tumor diameter over time after twice-weekly injection of anti-CD8, anti-NK1.1 neutralizing antibodies, or isotype control is displayed. One experiment out of two with 5 mice per group is shown. Data are means ± SEM. C, D, and H, Mann–Whitney test. I–M, Multiple t tests with FDR correction. F, Log-rank test. *, P < 0.05; **, P < 0.01.

Close modal

PTDSS1-dependent ePS shape macrophage activation by apoptotic cells via MERTK

PS exposed on dying cells is recognized by macrophages as an “eat me” signal (30). To determine whether the reduction in ePS after Ptdss1-KD translates into altered PS exposure during apoptosis, Ptdss1-KD PyMT cells and control PyMT cells were treated with staurosporine for up to 24 hours. A strong reduction in PS exposure on Ptdss1-knockdown cells 24 hours after apoptosis induction was observed, whereas cell death induction indicated by cell shrinkage, propidium iodide staining, as well as phagocytosis of Ptdss1-KD cells by IC-21 macrophages was unaltered (Fig. 4A and B; Supplementary Figs. S5A–S5D). These data suggested a major contribution of ePS to PS exposure. Next, a FACS-FRET assay to determine Annexin V binding to PS exposed on apoptotic cells was established. Jurkat T cells were incubated with NBD-labeled PS for 2 hours prior to induction of apoptosis. Exposed PS was detected with Annexin V-V450, allowing efficient energy transfer to NBD (Supplementary Fig. S5E). Importantly, apoptotic Jurkat cells produced a stronger FRET signal than viable Jurkat cells (Supplementary Fig. S5F). With this setup, a competition assay was performed where unlabeled lyso-PS or lyso-ePS was added to Jurkat cells along with PS-NBD at different ratios. Hereby, lyso-ePS interfered with FRET activity more efficiently than lyso-PS (Fig. 4C), indicating that ePS species modulate PS exposure during apoptosis.

Figure 4.

PTDSS1-dependent PS signals through MERTK on macrophages. A and B, PyMT cells lentivirally transduced with nontargeting ShRNA (Ctrl) or ShRNA-targeting Ptdss1 (Ptdss1-KD) were treated with 0.5 µg/mL staurosporine (Sts) for the times indicated. A, Representative FACS plots show cell size (FSC) and PS exposure (Annexin V-APC). B, Mean fluorescence intensity of Annexin V-APC relative to unstimulated control cells is displayed (n = 3). C, Jurkat T cells were prelabeled with NBD-PS for 2 hours with or without addition of lyso-PS or lyso-ePS at the ratios indicated, followed by staurosporine treatment for 3 hours. Cells were stained with Annexin V-V450 and FRET from V450 to NBD and was calculated. D–I, IC-21 murine macrophages were controls or stimulated with supernatants of viable (VCM) or apoptotic (ACM) control (Ctrl) or Ptdss1-knockdown (Ptdss1-KD) PyMT cells, with or without the addition of lyso-ePS or MERTK inhibitor. Tnfa mRNA expression after 6 hours of stimulation (D, G, and I), expression of macrophage markers at mRNA level after 6 hours (E), TNFα protein secretion determined by cytometric bead array after 24 hours (F), and expression of PS receptors at mRNA level (H) are shown. Data are means ± SEM. B, Two-way ANOVA with Bonferroni correction. C, D, F, G, and I, One-way ANOVA with FDR correction. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 4.

PTDSS1-dependent PS signals through MERTK on macrophages. A and B, PyMT cells lentivirally transduced with nontargeting ShRNA (Ctrl) or ShRNA-targeting Ptdss1 (Ptdss1-KD) were treated with 0.5 µg/mL staurosporine (Sts) for the times indicated. A, Representative FACS plots show cell size (FSC) and PS exposure (Annexin V-APC). B, Mean fluorescence intensity of Annexin V-APC relative to unstimulated control cells is displayed (n = 3). C, Jurkat T cells were prelabeled with NBD-PS for 2 hours with or without addition of lyso-PS or lyso-ePS at the ratios indicated, followed by staurosporine treatment for 3 hours. Cells were stained with Annexin V-V450 and FRET from V450 to NBD and was calculated. D–I, IC-21 murine macrophages were controls or stimulated with supernatants of viable (VCM) or apoptotic (ACM) control (Ctrl) or Ptdss1-knockdown (Ptdss1-KD) PyMT cells, with or without the addition of lyso-ePS or MERTK inhibitor. Tnfa mRNA expression after 6 hours of stimulation (D, G, and I), expression of macrophage markers at mRNA level after 6 hours (E), TNFα protein secretion determined by cytometric bead array after 24 hours (F), and expression of PS receptors at mRNA level (H) are shown. Data are means ± SEM. B, Two-way ANOVA with Bonferroni correction. C, D, F, G, and I, One-way ANOVA with FDR correction. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

Next, consequences of reduced PS exposure upon Ptdss1 knockdown were analyzed. IC-21 murine macrophages were stimulated with supernatants of living and apoptotic Ptdss1-KD and Ctrl PyMT cells. Treating IC-21 macrophages with supernatants of apoptotic but not viable PyMT cells suppressed Tnfa mRNA expression, which is a prominent feature of apoptotic cells (31). This did not occur when IC-21 macrophages received supernatants of Ptdss1-KD PyMT cells (Fig. 4D). Overall, supernatants from apoptotic Ptdss1-KDPyMT cells altered the expression profile of a number of macrophage activation markers compared with supernatants of Ctrl PyMT cells (Fig. 4E). TNFα expression was most consistently regulated, which was confirmed at protein level (Fig. 4F), and was therefore used as a read-out in further experiments. When lyso-ePS was pre-incubated with apoptotic PyMT cell supernatants to allow its interaction with apoptotic bodies, its addition to apoptotic Ctrl cell supernatants did not alter Tnfa suppression whereas addition to supernatants of apoptotic Ptdss1-KD cells rescued Tnfa suppression (Fig. 4G). Among PS receptors, IC-21 macrophages predominantly expressed the tyrosine kinase MERTK (Fig. 4H). Pharmacologic inhibition of MERTK prevented Tnfa suppression by supernatants of apoptotic Ctrl but not Ptdss1-KD cells (Fig. 4I). These data suggested that Ptdss1 promoted ePS-dependent PS exposure during apoptosis, which in turn modulated the macrophage phenotype via MERTK.

Tumor cell PTDSS1 and macrophage MERTK promote macrophage proliferation

To validate the role of MERTK in vivo, we used Csf1r-Cre Mertkf/f mice, in which Mertk is deleted in cells expressing the macrophage colony-stimulating factor receptor (CSF1R; ref. 13). Csf1r-Cre+ Mertkf/f and Csf1r-Cre Mertkf/f mice were orthotopically grafted with control or Ptdss1-KD PyMT cells and tumors were monitored for 5 weeks. MERTK ablation in macrophages reduced growth of control PyMT tumors similarly to the depletion of Ptdss1 in tumor cells, whereas no further reduction of Ptdss1-KD tumor growth was observed in Csf1r-Cre+ Mertkf/f mice (Fig. 5A). Both, Ptdss1-KD in tumor cells and MERTK-ablation in macrophages reduced TAM levels compared with Ctrl PyMT tumors (Fig. 5B), which was confirmed by multispectral imaging (Fig. 5C and D). Neither Ptdss1-KD in tumor cells nor MERTK-ablation in macrophages resulted in a clear M1 shift of M2 TAM (32). Although the M1 marker TNFα was upregulated the M1 marker IL12 and the M2 markers CD206 and ALOX15 were not affected (Fig. 5EH). Interestingly, reduced TAM infiltrates were even more apparent 3 weeks after tumor cell implantation when tumor volumes were similar, whereas NK-cell and CD8+ T-cell numbers and activation status were not significantly altered (Fig. 5I and J). Moreover, ePS levels were strongly reduced in Ptdss1-KD tumors 3 weeks after tumor implantation, which was blunted after 5 weeks (Fig. 5K and L). These data suggested that reduced ePS levels and TAM infiltrates preceded reduced tumor growth upon Ptdss1 KD.

Figure 5.

MERTK in macrophages supports tumor growth and macrophage accumulation. PyMT cells lentivirally transduced with nontargeting ShRNA (Ctrl) or shRNA-targeting Ptdss1 (Ptdss1-KD) were grafted into syngenic Csf1r-Cre Mertkf/f (WT), or Csf1r-Cre+ Mertkf/f (MertkΔMΦ) mice, and tumors were maintained for 5 (A–H and L) or 3 weeks (I–K). Tumor diameters over time (A) and TAM abundance (B) is shown. One experiments out of two with ≥4 mice per group. C–H, Tumor section were analyzed for expression of pan-Cytokeratin (pan-CK), CD163, CD206, ALOX15, IL12, and TNFα. Representative images (C) and the quantification of CD163+ macrophages (D) and macrophage subsets (E–H) within tumor stroma are shown. Data of seven fields of six (WT-Ctrl and WT-Ptdss1-KD) or four (MertkΔM) tumors. I and J, Tumor-infiltrating immune cells within total living cells (I), and perforin (PRF1) and IFNγ-expressing CD8+ T cells or NK cells (J) within the parent population 3 weeks after tumor cell injection determined by flow cytometry. K and L, Expression levels of ePS species 3 (K) or 5 (L) weeks after tumor cell injection are shown. Data are means ± SEM. A, Two-way ANOVA. B and D–H, One-way ANOVA with FDR correction. I–L, Multiple t tests with FDR correction. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 5.

MERTK in macrophages supports tumor growth and macrophage accumulation. PyMT cells lentivirally transduced with nontargeting ShRNA (Ctrl) or shRNA-targeting Ptdss1 (Ptdss1-KD) were grafted into syngenic Csf1r-Cre Mertkf/f (WT), or Csf1r-Cre+ Mertkf/f (MertkΔMΦ) mice, and tumors were maintained for 5 (A–H and L) or 3 weeks (I–K). Tumor diameters over time (A) and TAM abundance (B) is shown. One experiments out of two with ≥4 mice per group. C–H, Tumor section were analyzed for expression of pan-Cytokeratin (pan-CK), CD163, CD206, ALOX15, IL12, and TNFα. Representative images (C) and the quantification of CD163+ macrophages (D) and macrophage subsets (E–H) within tumor stroma are shown. Data of seven fields of six (WT-Ctrl and WT-Ptdss1-KD) or four (MertkΔM) tumors. I and J, Tumor-infiltrating immune cells within total living cells (I), and perforin (PRF1) and IFNγ-expressing CD8+ T cells or NK cells (J) within the parent population 3 weeks after tumor cell injection determined by flow cytometry. K and L, Expression levels of ePS species 3 (K) or 5 (L) weeks after tumor cell injection are shown. Data are means ± SEM. A, Two-way ANOVA. B and D–H, One-way ANOVA with FDR correction. I–L, Multiple t tests with FDR correction. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

To determine mechanisms underlying the decrease in TAM abundance, TAMs were isolated from Ctrl and Ptdss1-KD tumors growing in Csf1r-Cre Mertkf/f or Csf1r-Cre+ Mertkf/f mice, and their transcriptomes were analyzed by transcriptome sequencing. When comparing genes with a log2 FC ≥ 0.5 between macrophages from Ctrl versus Ptdss1-KD tumors, and between macrophages from tumors grown in Csf1r-Cre Mertkf/f versus Csf1r-Cre+ Mertkf/f mice, there was a clear overlap of upregulated or downregulated genes (Fig. 6A). A shared gene signature of 580 differentially expressed genes was found (Fig. 6B; Supplementary Table S2). Among these genes was Tnfa, supporting the in vitro data obtained with IC-21 macrophages (Fig. 4) and the TNFα protein expression in macrophages (Fig. 5H). GSEA with this gene signature revealed a number of pathways altered in macrophages from Ptdss1-KD tumors and Mertk-deficient TAMs compared with TAMs from control tumors. Pathways enriched in Ctrl TAM indicated increased cell-cycle progression (Fig. 6C; Supplementary Table S3). Pathways enriched in the other two TAM populations were more diverse, indicating increased defense response and extracellular matrix organization (Supplementary Table S3), whereas again no clear shift in M1/M2 polarization was found. To validate the impact of tumor cell PTDSS1 and macrophage MERTK on macrophage proliferation, IC-21 macrophages were stimulated with PyMT cell supernatants for 72 hours. Macrophages receiving medium from viable PyMT cells did proliferate more than control cells, but supernatants from apoptotic PyMT cells slightly reduced IC-21 macrophage proliferation. However, proliferation was strongly reduced when macrophages received supernatants from either apoptotic Ptdss1-KD PyMT cells or control PyMT cells supplemented with a MERTK inhibitor (Fig. 6D and E). Macrophage numbers in the latter groups were reduced below the initial seeding density, likely due to increased apoptosis (Fig. 6F). Reduced proliferation was confirmed by validation of proliferation-associated genes in the shared TAM gene signature, namely Ccni and Pole (Supplementary Fig. S6). These genes were downregulated in IC-21 cells treated with apoptotic Ptdss1-KD cell supernatants and supernatants of apoptotic Ctrl PyMT cells combined with a MERTK inhibitor. Addition of lyso-ePS but not lyso-PS restored expression of Ccni and Pole in IC-21 cells treated with supernatants of apoptotic Ptdss1-KD PyMT cells. Genes involved in macrophage activation such as Adgre5 (CD97) or Timp1 were regulated inversely (Supplementary Fig. S6). These data suggest that PTDSS1-derived ePS promote macrophage proliferation via MERTK, which appears to be involved in regulating TAM levels and tumor progression. Promoting angiogenesis and tumor cell proliferation are features of tumor-supporting macrophages. Although the amount of endothelial cells was unaltered in tumors with ablation of tumor cell PTDSS1 or macrophage MERTK (Supplementary Fig. S4B), Ki67 expression, indicating proliferation, was decreased when compared with Ctrl tumors (Fig. 6G and H). Moreover, PyMT cells cultured with supernatants from IC-21 cells showed increased proliferation indicated by Ki67 expression, when being pretreated with apoptotic PyMT cells, but not Ptdss1-ablated apoptotic cells or when a MERTK inhibitor was used (Fig. 6I). Thus, ePS/MERTK signaling in TAM may support tumor cell proliferation.

Figure 6.

Tumor cell PTDSS1 and macrophage MERTK promote macrophage proliferation. A–C, Macrophages were FACS-sorted from mammary tumors, and transcriptomes were analyzed by whole transcriptome mRNA sequencing. A, Venn diagram compared upregulated or downregulated genes of macrophages isolated from Ptdss1-KD tumors with macrophages from MertkΔMΦ Ctrl tumors (each compared with macrophages from Ctrl tumors). B, Heatmap indicates shared gene signature between macrophages from Ptdss1-KD tumors and MertkΔMΦ TAMs. C, GSEA plots of representative enriched pathways in macrophages from WT Ctrl tumors compared with both, macrophages from Ptdss1-KD tumors and MertkΔMΦ TAMs. D–G, IC-21 murine macrophages were controls or stimulated with supernatants of viable (VCM) or apoptotic (ACM) control (Ctrl) or Ptdss1-knockdown (Ptdss1-KD) PyMT cells, with or without the addition of MERTK inhibitor for 72 hours. Absolute cell numbers (D), mean fluorescence intensity of the proliferation dye eFluor670 (E), and cell viability determined by Annexin V/PI assay (F), each determined via flow cytometry, are shown. G and H, Tumor sections were analyzed for expression of Ki67, pan-Cytokeratin (pan-CK), and vimentin (Vim). The quantification of Ki67+ tumor cells (G) and representative images (H) are shown. Data of seven fields of each six (WT-Ctrl and WT-Ptdss1-KD) or four (MertkΔM) tumors. I, IC-21 macrophage supernatants (macrophage CM) were added to tumor cells for 24 hours and Ki67 mRNA expression was determined. Data are means ± SEM. D, E, G, and I, One-way ANOVA with FDR correction. F, Two-way ANOVA with Bonferroni correction. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 6.

Tumor cell PTDSS1 and macrophage MERTK promote macrophage proliferation. A–C, Macrophages were FACS-sorted from mammary tumors, and transcriptomes were analyzed by whole transcriptome mRNA sequencing. A, Venn diagram compared upregulated or downregulated genes of macrophages isolated from Ptdss1-KD tumors with macrophages from MertkΔMΦ Ctrl tumors (each compared with macrophages from Ctrl tumors). B, Heatmap indicates shared gene signature between macrophages from Ptdss1-KD tumors and MertkΔMΦ TAMs. C, GSEA plots of representative enriched pathways in macrophages from WT Ctrl tumors compared with both, macrophages from Ptdss1-KD tumors and MertkΔMΦ TAMs. D–G, IC-21 murine macrophages were controls or stimulated with supernatants of viable (VCM) or apoptotic (ACM) control (Ctrl) or Ptdss1-knockdown (Ptdss1-KD) PyMT cells, with or without the addition of MERTK inhibitor for 72 hours. Absolute cell numbers (D), mean fluorescence intensity of the proliferation dye eFluor670 (E), and cell viability determined by Annexin V/PI assay (F), each determined via flow cytometry, are shown. G and H, Tumor sections were analyzed for expression of Ki67, pan-Cytokeratin (pan-CK), and vimentin (Vim). The quantification of Ki67+ tumor cells (G) and representative images (H) are shown. Data of seven fields of each six (WT-Ctrl and WT-Ptdss1-KD) or four (MertkΔM) tumors. I, IC-21 macrophage supernatants (macrophage CM) were added to tumor cells for 24 hours and Ki67 mRNA expression was determined. Data are means ± SEM. D, E, G, and I, One-way ANOVA with FDR correction. F, Two-way ANOVA with Bonferroni correction. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

PTDSS1 expression correlates with survival in patients with cancer

Analyzing the relevance of our findings in human cancer, a negative association of PTDSS1 with breast cancer survival suggested in the PRECOG database (33), was confirmed in the METABRIC cohort (21) after correction for noncancer-related causes of death (1,484 patients remaining; Fig. 7A). Among enzymes potentially involved in ePS synthesis, only fatty acid reductase 2 (FAR2; ref. 34), was also associated with negative patient survival (Fig. 7A). A closer look into molecular subtypes of BC (35) revealed a particular association of PTDSS1 expression with survival of basal-like patients with breast cancer (Fig. 7B), corresponding to high PTDSS1 expression in basal-like breast tumors (Fig. 7C). Correlation analyses furthermore revealed a positive association of PTDSS1 with tumor grade, but a negative association with ER and PR expression (Fig. 7D). The negative association of PTDSS1 with patient survival was also observed in a data set of 295 patients with invasive mammary cancer (Fig. 7E; ref. 22), where PTDSS1 expression was higher in patients with metastasis (Fig. 7F). Also at protein level in tissue microarrays, PTDSS1 correlated with breast cancer patient survival (Fig. 7G and H), where PTDSS1 expression was observed in tumor but not stromal cells. Correlation of PTDSS1 with survival was specifically observed for PR+ and TNBC (Fig. 7I), corroborating the data of the METABRIC cohort, as a majority of TNBC are of a basal-like subtype (36). Moreover, PTDSS1 expression correlated with metastasis (Fig. 7J). The negative correlation of PTDSS1 with survival was also observed in the lung adenocarcinoma data set of TCGA program (Fig. 7K). Overall, these data suggest that PTDSS1 expression promotes cancer progression and mortality in humans. Next, PTDSS1 expression was correlated with macrophage markers. A correlation between the expression of the macrophage marker CD163 and PTDSS1 was observed in the METABRIC cohort. However, this correlation did not follow the same pattern as the association of PTDSS1 with survival in breast cancer subtypes (Fig. 7L), indicating that the tumor-promoting effect of PTDSS1 may transcend affecting macrophages in humans. Nevertheless, CD163+ macrophages at the cellular level were highly correlated with PTDSS1 expression in invasive mammary carcinoma tissue microarrays, and we observed a slightly pronounced correlation of PTDSS1 expression with macrophage infiltrates in TNBC (Fig. 7M).

Figure 7.

PTDSS1 expression correlates with survival. A–F, Association of PS and ether lipid metabolizing enzymes with clinical parameters in publicly available breast cancer data sets. Correlation of PS and ether lipid metabolizing enzymes with survival (A), correlation of PTDSS1 with survival (B), PTDSS1 expression in breast cancer subtypes (C), and correlation of PTDSS1 expression with clinical parameters in the METABRIC cohort (D). Parameters used as binary categorical variables are in bold; continuous variables in light script. Correlation of PTDSS1 expression with survival (E) and metastasis (F) in the van de Vijver data set. G–J, PTDSS1 expression and clinical parameters in human breast cancer tissue microarrays. G, Representative microscopy images show CD163, pan-Cytokeratin (Pan-CK), and PTDSS1 expression in mammary carcinoma cores; nuclei stained with DAPI. Scale bars, 100 µm. Correlation of PTDSS1 expression with survival of all patients (H) and with survival of patients with estrogen receptor (ER), progesterone receptor (PR), HER2-expressing or triple-negative breast tumors (TNBC; I). J, Correlation of PTDSS1 expression with clinical parameters. Parameters used as binary categorical variables are in bold; continuous variables in light script.K, Correlation of PTDSS1 expression with patient survival in the TCGA lung adenocarcinoma data set. L, Correlation of PTDSS1 and macrophage marker CD163 expression in breast cancer subtypes in the METABRIC dataset. M, PTDSS1 and CD163 expression in tissue microarrays. Representative microscopy images show cells expressing CD163 and PTDSS1; nuclei stained with DAPI. Scale bars, 100 µm. The graph shows PTDSS1 versus CD163 expression in breast cancer subtypes. A, B, E, H, I, and K, Log-rank test. D, J, L, and M, Spearman correlation coefficient when comparing continuous variables (D, J, L, and M) and point-biserial correlation coefficient when comparing continuous with binary categorical variables are shown (D and J). Significant correlations (P < 0.05) are in bold. F, Unpaired two-tailed Student t test; ***, P < 0.001.

Figure 7.

PTDSS1 expression correlates with survival. A–F, Association of PS and ether lipid metabolizing enzymes with clinical parameters in publicly available breast cancer data sets. Correlation of PS and ether lipid metabolizing enzymes with survival (A), correlation of PTDSS1 with survival (B), PTDSS1 expression in breast cancer subtypes (C), and correlation of PTDSS1 expression with clinical parameters in the METABRIC cohort (D). Parameters used as binary categorical variables are in bold; continuous variables in light script. Correlation of PTDSS1 expression with survival (E) and metastasis (F) in the van de Vijver data set. G–J, PTDSS1 expression and clinical parameters in human breast cancer tissue microarrays. G, Representative microscopy images show CD163, pan-Cytokeratin (Pan-CK), and PTDSS1 expression in mammary carcinoma cores; nuclei stained with DAPI. Scale bars, 100 µm. Correlation of PTDSS1 expression with survival of all patients (H) and with survival of patients with estrogen receptor (ER), progesterone receptor (PR), HER2-expressing or triple-negative breast tumors (TNBC; I). J, Correlation of PTDSS1 expression with clinical parameters. Parameters used as binary categorical variables are in bold; continuous variables in light script.K, Correlation of PTDSS1 expression with patient survival in the TCGA lung adenocarcinoma data set. L, Correlation of PTDSS1 and macrophage marker CD163 expression in breast cancer subtypes in the METABRIC dataset. M, PTDSS1 and CD163 expression in tissue microarrays. Representative microscopy images show cells expressing CD163 and PTDSS1; nuclei stained with DAPI. Scale bars, 100 µm. The graph shows PTDSS1 versus CD163 expression in breast cancer subtypes. A, B, E, H, I, and K, Log-rank test. D, J, L, and M, Spearman correlation coefficient when comparing continuous variables (D, J, L, and M) and point-biserial correlation coefficient when comparing continuous with binary categorical variables are shown (D and J). Significant correlations (P < 0.05) are in bold. F, Unpaired two-tailed Student t test; ***, P < 0.001.

Close modal

Serine production as well as ether lipid content were previously associated with cancer development (12, 37). Serine exerts multiple functions in metabolism, including one-carbon metabolism and the de novo synthesis of sphingolipids, but if its incorporation into phospholipids is involved in tumor growth was unclear (38). PS constitutes only 3% to 10% of total phospholipids (7.8% in mouse mammary tissue in our study). However, in PyMT tumors, PS levels increased to 13.1% of total phospholipids. This may have significant consequences for tumor cells themselves, as PS is essential for intracellular proliferation and survival signaling by interacting with Ras-, Rho-GTPases, and protein kinases B/C (39). Along this line, PTDSS1 was connected to CHO cell proliferation in vitro (40). This did not occur in PyMT cells, probably due to incomplete PTDSS1-KD or compensation by PTDSS2. Accordingly, 85% reduced serine-exchange capability in Ptdss1–/– mice did not alter the PS content due to compensation by PTDSS2, whereas only deletion of both synthases was lethal (40).

PTDSS1-KD in tumor cells mainly reduced ePS. ePS are likely rare in tissues, probably below 1% of total phospholipids (41). Nevertheless, reduction of ePS by PTDSS1-KD affected PS exposure during apoptosis and altered macrophage gene expression in response to dying cells. It remains unclear how ePS species affect PS exposure. Ether lipids, and 1-O-alkenyl ether lipids in particular, were proposed as cellular antioxidants (34). Oxidation of PS is a requirement for its exposure during apoptosis (42). However, we used PS(O:18:0/0:0) in our competition assay, which is probably not a potent ROS scavenger. An alternative explanation might be the absence of the carbonyl oxygen at the sn-1 position in ether-linked versus ePS, which increases intermolecular hydrogen bonding and, thus, membrane rigidity (34), facilitating cellular signaling by promoting membrane microdomain formation. Detailed mechanistic investigations to understand the specific contribution of ePS to PS exposure are currently hampered by commercial unavailability of labeled ePS species.

PS is recognized by a number of receptors including the TAM receptors TYRO3, AXL, and MERTK (43). MERTK is a largely macrophage-specific receptor (44), indicating that Mertk deletion in Csf1r-Cre+ Mertkf/f mice selectively affects macrophages, even though CSF1R is also expressed by other immune cells. MERTK signaling in macrophages is relevant in inflammatory disease and cancer. MERTK-expressing macrophages connected resolution of inflammation to tissue repair after cardiac infarction (45), and MERTK was required for pro-resolving lipid mediators synthesis (46). Efferocytosis through MERTK altered macrophage phenotypes to promote tumor progression and metastasis in breast cancer (47). Blocking MERTK with antibodies triggered antitumor T-cell responses, especially after combination with immune checkpoint blockade (48). In our study, targeting PTDSS1, which affected MERTK signaling also increased cytotoxic lymphocyte activation, and we observed trends towards increased CD8+ T cells and NK cells, which has been observed before in a Mertk-ablated background (49). Neutralization experiments suggested a minor impact of cytotoxic immune cells in our model. Rather, MERTK signaling promoted macrophage proliferation, which in turn supported tumor cell proliferation. Proliferation is a crucial feature of macrophage expansion in mammary tumors (50). An impact of MERTK on macrophage proliferation was previously observed during helminth infection (13), although mechanisms of PS/MERTK-dependent macrophage proliferation remain to be determined.

PTDSS1 may be a means to interfere with tumor-promoting macrophage responses, because reducing PS surface exposure did not prevent efferocytosis per se, thus avoiding the autoimmune side-effects (51). Cell surface exposed PS has already been clinically targeted with an mAb, Bavituximab. Results of a phase III clinical trial in non–small cell lung cancer were discouraging, even though combining Bavituximab with immune checkpoint blockade appeared promising (52). Bavituximab was found in complex with the tumor vasculature rather than tumor cells (51) and binding to the vasculature may prevent its potential to interrupt the impact of dying tumor cells on TAMs. This limitation would likely not apply to PTDSS1-targeting agents. Therefore, identifying pharmacologic compounds targeting PTDSS1 may be of interest for tumor therapy. Our data suggest that this may be of particular relevance for TNBC/basal tumors rather than breast cancer per se. Interestingly, immunotherapy appears promising in TNBC, among others due to a high lymphocyte infiltrate (53). Because targeting PTDSS1 increased cytotoxic lymphocyte activation, future studies addressing a potential synergism of immune checkpoint blockade with PTDSS1 inhibition may be of interest.

A. Descot reports grants from European Research Council during the conduct of the study and nonfinancial support from Bergenbio outside the submitted work. I. Fleming reports grants from Deutsche Forschungsgemeinschaft during the conduct of the study. H. Stark reports grants from GRK2158 outside the submitted work and also has a patent for EP2505589 issued. H. Medyouf reports grants from ERC Starting, Hessen State Ministry for Higher Education, Research and the Arts, and DFG SPP2084 during the conduct of the study and grants and personal fees from BergenBio Inc. outside the submitted work. No disclosures were reported by the other authors.

D. Sekar: Conceptualization, formal analysis, investigation, methodology, writing–original draft, writing–review and editing. C. Dillmann: Formal analysis, investigation, visualization, methodology, writing–review and editing. E. Sirait-Fischer: Formal analysis, investigation, methodology, writing–review and editing. A.F. Fink: Investigation, methodology, writing–review and editing. A. Zivkovic: Investigation, methodology, writing–review and editing. N. Baum: Investigation, methodology, writing–review and editing. E. Strack: Investigation, writing–review and editing. S. Klatt: Formal analysis, investigation, methodology, writing–review and editing. S. Zukunft: Investigation, methodology, writing–review and editing. S. Wallner: Investigation, methodology, writing–review and editing. A. Descot: Resources, investigation, writing–review and editing. C. Olesch: Investigation, writing–review and editing. P. da Silva: Investigation, writing–review and editing. A. von Knethen: Resources, methodology, writing–review and editing. T. Schmid: Formal analysis, methodology, writing–review and editing. S. Groesch: Resources, writing–review and editing. R. Savai: Resources, writing–review and editing. N. Ferreirós: Investigation, methodology, writing–review and editing. I. Fleming: Resources, supervision, writing–review and editing. S. Ghosh: Resources, writing–review and editing. C.V. Rothlin: Resources, writing–review and editing. H. Stark: Resources, supervision, methodology, writing–review and editing. H. Medyouf: Resources, supervision, writing–review and editing. B. Brüne: Resources, writing–original draft, writing–review and editing. A. Weigert: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing.

The authors thank Margarete Mijatovic, Bettina Wenzel, and Praveen Mathoor for excellent technical assistance. This work was supported by Deutsche Forschungsgemeinschaft (SFB 1039, TP B04, and B06; GRK 2158, GRK 2336, TP1, and 6), Deutsche Krebshilfe (70114051), Wilhelm-Sander Foundation (2019.082.01), the LOEWE Center for Translational Medicine and Pharmacology, and the LOEWE Center Frankfurt Cancer Institute (FCI), funded by the Hessen State Ministry for Higher Education, Research and the Arts.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
ICGC/TCGA Pan-Cancer Analysis of Whole Genomes Consortium
.
Pan-cancer analysis of whole genomes
.
Nature
2020
;
578
:
82
93
.
2.
Letai
A
.
Functional precision cancer medicine-moving beyond pure genomics
.
Nat Med
2017
;
23
:
1028
35
.
3.
Baenke
F
,
Peck
B
,
Miess
H
,
Schulze
A
.
Hooked on fat: the role of lipid synthesis in cancer metabolism and tumour development
.
Dis Model Mech
2013
;
6
:
1353
63
.
4.
Wang
D
,
Dubois
RN
.
Eicosanoids and cancer
.
Nat Rev Cancer
2010
;
10
:
181
93
.
5.
Zalba
S
,
Ten Hagen
TL
.
Cell membrane modulation as adjuvant in cancer therapy
.
Cancer Treat Rev
2017
;
52
:
48
57
.
6.
Rivel
T
,
Ramseyer
C
,
Yesylevskyy
S
.
The asymmetry of plasma membranes and their cholesterol content influence the uptake of cisplatin
.
Sci Rep
2019
;
9
:
5627
.
7.
Utsugi
T
,
Schroit
AJ
,
Connor
J
,
Bucana
CD
,
Fidler
IJ
.
Elevated expression of phosphatidylserine in the outer membrane leaflet of human tumor cells and recognition by activated human blood monocytes
.
Cancer Res
1991
;
51
:
3062
6
.
8.
Kirszberg
C
,
Lima
LG
,
Da Silva de Oliveira
A
,
Pickering
W
,
Gray
E
,
Barrowcliffe
TW
, et al
.
Simultaneous tissue factor expression and phosphatidylserine exposure account for the highly procoagulant pattern of melanoma cell lines
.
Melanoma Res
2009
;
19
:
301
8
.
9.
Terabe
M
,
Berzofsky
JA
.
The role of NKT cells in tumor immunity
.
Adv Cancer Res
2008
;
101
:
277
348
.
10.
Merrill
AH
Jr
,
Stokes
TH
,
Momin
A
,
Park
H
,
Portz
BJ
,
Kelly
S
, et al
.
Sphingolipidomics: a valuable tool for understanding the roles of sphingolipids in biology and disease
.
J Lipid Res
2009
;
50
:
S97
102
.
11.
Kiebish
MA
,
Han
X
,
Cheng
H
,
Chuang
JH
,
Seyfried
TN
.
Cardiolipin and electron transport chain abnormalities in mouse brain tumor mitochondria: lipidomic evidence supporting the Warburg theory of cancer
.
J Lipid Res
2008
;
49
:
2545
56
.
12.
Benjamin
DI
,
Cozzo
A
,
Ji
X
,
Roberts
LS
,
Louie
SM
,
Mulvihill
MM
, et al
.
Ether lipid generating enzyme AGPS alters the balance of structural and signaling lipids to fuel cancer pathogenicity
.
Proc Natl Acad Sci U S A
2013
;
110
:
14912
7
.
13.
Bosurgi
L
,
Cao
YG
,
Cabeza-Cabrerizo
M
,
Tucci
A
,
Hughes
LD
,
Kong
Y
, et al
.
Macrophage function in tissue repair and remodeling requires IL-4 or IL-13 with apoptotic cells
.
Science
2017
;
356
:
1072
6
.
14.
Chang
DH
,
Deng
H
,
Matthews
P
,
Krasovsky
J
,
Ragupathi
G
,
Spisek
R
, et al
.
Inflammation-associated lysophospholipids as ligands for CD1d-restricted T cells in human cancer
.
Blood
2008
;
112
:
1308
16
.
15.
Liebisch
G
,
Lieser
B
,
Rathenberg
J
,
Drobnik
W
,
Schmitz
G
.
High-throughput quantification of phosphatidylcholine and sphingomyelin by electrospray ionization tandem mass spectrometry coupled with isotope correction algorithm
.
Biochim Biophys Acta
2004
;
1686
:
108
17
.
16.
Koch
S
,
Schollmeyer
D
,
Lowe
H
,
Kunz
H
.
C-Glycosyl amino acids through hydroboration-cross-coupling of exo-glycals and their application in automated solid-phase synthesis
.
Chemistry
2013
;
19
:
7020
41
.
17.
Bandyopadhyay
S
,
Bong
D
.
Synthesis of trifunctional phosphatidylserine probes for identification of novel lipid-binding proteins
.
Abstr Pap Am Chem S
2011
;
241
.
18.
Ni
GH
,
Li
ZY
,
Liang
KJ
,
Wu
T
,
De Libero
G
,
Xia
CF
.
Synthesis and evaluation of immunostimulant plasmalogen lysophosphatidylethanolamine and analogues for natural killer T cells
.
Bioorgan Med Chem
2014
;
22
:
2966
73
.
19.
Xu
Y
,
Qian
L
,
Prestwich
GD
.
Synthesis of monofluorinated analogues of lysophosphatidic acid
.
J Org Chem
2003
;
68
:
5320
.
20.
Hoadley
KA
,
Yau
C
,
Hinoue
T
,
Wolf
DM
,
Lazar
AJ
,
Drill
E
, et al
.
Cell-of-origin patterns dominate the molecular classification of 10,000 tumors from 33 types of cancer
.
Cell
2018
;
173
:
291
304
.
21.
Curtis
C
,
Shah
SP
,
Chin
SF
,
Turashvili
G
,
Rueda
OM
,
Dunning
MJ
, et al
.
The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups
.
Nature
2012
;
486
:
346
52
.
22.
van de Vijver
MJ
,
He
YD
,
van't Veer
LJ
,
Dai
H
,
Hart
AA
,
Voskuil
DW
, et al
.
A gene-expression signature as a predictor of survival in breast cancer
.
N Engl J Med
2002
;
347
:
1999
2009
.
23.
Olesch
C
,
Sirait-Fischer
E
,
Berkefeld
M
,
Fink
AF
,
Susen
RM
,
Ritter
B
, et al
.
S1PR4 ablation reduces tumor growth and improves chemotherapy via CD8+ T cell expansion
.
J Clin Invest
2020
;
130
:
5461
76
.
24.
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
.
25.
Dixson
DD
,
Yu
VY
,
Doyle
RP
.
Recombinant expression of His-tagged saposin B and pH-dependent binding to the lipid coenzyme Q10
.
Anal Biochem
2011
;
419
:
145
52
.
26.
Ujlaky-Nagy
L
,
Nagy
P
,
Szollosi
J
,
Vereb
G
.
Flow cytometric FRET analysis of protein interactions
.
Methods Mol Biol
2018
;
1678
:
393
419
.
27.
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
.
28.
Reich
M
,
Liefeld
T
,
Gould
J
,
Lerner
J
,
Tamayo
P
,
Mesirov
JP
.
GenePattern 2.0
.
Nat Genet
2006
;
38
:
500
1
.
29.
Forbes
SA
,
Bindal
N
,
Bamford
S
,
Cole
C
,
Kok
CY
,
Beare
D
, et al
.
COSMIC: mining complete cancer genomes in the catalogue of somatic mutations in cancer
.
Nucleic Acids Res
2011
;
39
:
D945
50
.
30.
Nagata
S
.
Apoptosis and clearance of apoptotic cells
.
Annu Rev Immunol
2018
;
36
:
489
517
.
31.
Voll
RE
,
Herrmann
M
,
Roth
EA
,
Stach
C
,
Kalden
JR
,
Girkontaite
I
.
Immunosuppressive effects of apoptotic cells
.
Nature
1997
;
390
:
350
1
.
32.
Sica
A
,
Mantovani
A
.
Macrophage plasticity and polarization: in vivo veritas
.
J Clin Invest
2012
;
122
:
787
95
.
33.
Gentles
AJ
,
Newman
AM
,
Liu
CL
,
Bratman
SV
,
Feng
W
,
Kim
D
, et al
.
The prognostic landscape of genes and infiltrating immune cells across human cancers
.
Nat Med
2015
;
21
:
938
45
.
34.
Dean
JM
,
Lodhi
IJ
.
Structural and functional roles of ether lipids
.
Protein Cell
2018
;
9
:
196
206
.
35.
Sørlie
T
,
Tibshirani
R
,
Parker
J
,
Hastie
T
,
Marron
JS
,
Nobel
A
, et al
.
Repeated observation of breast tumor subtypes in independent gene expression data sets
.
Proc Natl Acad Sci
2003
;
100
:
8418
23
.
36.
Badve
S
,
Dabbs
DJ
,
Schnitt
SJ
,
Baehner
FL
,
Decker
T
,
Eusebi
V
, et al
.
Basal-like and triple-negative breast cancers: a critical review with an emphasis on the implications for pathologists and oncologists
.
Mod Pathol
2011
;
24
:
157
67
.
37.
Possemato
R
,
Marks
KM
,
Shaul
YD
,
Pacold
ME
,
Kim
D
,
Birsoy
K
, et al
.
Functional genomics reveal that the serine synthesis pathway is essential in breast cancer
.
Nature
2011
;
476
:
346
50
.
38.
Mattaini
KR
,
Sullivan
MR
,
Vander Heiden
MG
.
The importance of serine metabolism in cancer
.
J Cell Biol
2016
;
214
:
249
57
.
39.
Sousa
SB
,
Jenkins
D
,
Chanudet
E
,
Tasseva
G
,
Ishida
M
,
Anderson
G
, et al
.
Gain-of-function mutations in the phosphatidylserine synthase 1 (PTDSS1) gene cause Lenz-Majewski syndrome
.
Nat Genet
2014
;
46
:
70
6
.
40.
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
.
41.
Kim
HY
,
Huang
BX
,
Spector
AA
.
Phosphatidylserine in the brain: metabolism and function
.
Prog Lipid Res
2014
;
56
:
1
18
.
42.
Kagan
VE
,
Gleiss
B
,
Tyurina
YY
,
Tyurin
VA
,
Elenstrom-Magnusson
C
,
Liu
SX
, et al
.
A role for oxidative stress in apoptosis: oxidation and externalization of phosphatidylserine is required for macrophage clearance of cells undergoing Fas-mediated apoptosis
.
J Immunol
2002
;
169
:
487
99
.
43.
Moller-Tank
S
,
Maury
W
.
Phosphatidylserine receptors: enhancers of enveloped virus entry and infection
.
Virology
2014
;
468–470
:
565
80
.
44.
Gautier
EL
,
Shay
T
,
Miller
J
,
Greter
M
,
Jakubzick
C
,
Ivanov
S
, et al
.
Gene-expression profiles and transcriptional regulatory pathways that underlie the identity and diversity of mouse tissue macrophages
.
Nat Immunol
2012
;
13
:
1118
28
.
45.
Wan
E
,
Yeap
XY
,
Dehn
S
,
Terry
R
,
Novak
M
,
Zhang
S
, et al
.
Enhanced efferocytosis of apoptotic cardiomyocytes through myeloid-epithelial-reproductive tyrosine kinase links acute inflammation resolution to cardiac repair after infarction
.
Circ Res
2013
;
113
:
1004
12
.
46.
Cai
B
,
Kasikara
C
,
Doran
AC
,
Ramakrishnan
R
,
Birge
RB
,
Tabas
I
.
MerTK signaling in macrophages promotes the synthesis of inflammation resolution mediators by suppressing CaMKII activity
.
Sci Signal
2018
;
11
:
eaar3721
.
47.
Stanford
JC
,
Young
C
,
Hicks
D
,
Owens
P
,
Williams
A
,
Vaught
DB
, et al
.
Efferocytosis produces a prometastatic landscape during postpartum mammary gland involution
.
J Clin Invest
2014
;
124
:
4737
52
.
48.
Zhou
Y
,
Fei
M
,
Zhang
G
,
Liang
WC
,
Lin
W
,
Wu
Y
, et al
.
Blockade of the phagocytic receptor MerTK on tumor-associated macrophages enhances P2×7R-dependent STING activation by tumor-derived cGAMP
.
Immunity
2020
;
52
:
357
73
.
49.
Cook
RS
,
Jacobsen
KM
,
Wofford
AM
,
DeRyckere
D
,
Stanford
J
,
Prieto
AL
, et al
.
MerTK inhibition in tumor leukocytes decreases tumor growth and metastasis
.
J Clin Invest
2013
;
123
:
3231
42
.
50.
Franklin
RA
,
Liao
W
,
Sarkar
A
,
Kim
MV
,
Bivona
MR
,
Liu
K
, et al
.
The cellular and molecular origin of tumor-associated macrophages
.
Science
2014
;
344
:
921
5
.
51.
Birge
RB
,
Boeltz
S
,
Kumar
S
,
Carlson
J
,
Wanderley
J
,
Calianese
D
, et al
.
Phosphatidylserine is a global immunosuppressive signal in efferocytosis, infectious disease, and cancer
.
Cell Death Differ
2016
;
23
:
962
78
.
52.
Gerber
DE
,
Horn
L
,
Boyer
M
,
Sanborn
R
,
Natale
R
,
Palmero
R
, et al
.
Randomized phase III study of docetaxel plus bavituximab in previously treated advanced non-squamous non-small-cell lung cancer
.
Ann Oncol
2018
;
29
:
1548
53
.
53.
Thomas
R
,
Al-Khadairi
G
,
Decock
J
.
Immune checkpoint inhibitors in triple negative breast cancer treatment: promising future prospects
.
Front Oncol
2020
;
10
:
600573
.

Supplementary data