Plasmalogen-Mediated Activation of GPCR21 Regulates Cytolytic Activity of NK Cells against the Target Cells

In elderly subjects, reduced activity of NK cells was associated with the decline in immune responses to viral infection and tumor growth (1, 2). Recent studies also showed that aging is correlated with the mortality rate of coronavirus disease 2019 (3), suggesting that age-related changes of substances in our body could be a cause of reduced activity of NK cells. Although our knowledge remains limited in identifying age-related factors that regulate NK cell activity, several studies have shown a reduced blood level of lipids, including special phospholipids called plasmalogens (Pls), in older subjects (4–6). From this evidence, we investigated whether there is any functional relationship between Pls and NK cell activity.

We have previously reported that Pls are reduced by neuroinflammation, aging, and stress in mice (7, 8). In addition, we have also found that Pls have neuroprotective effects and can induce cellular signaling such as ERK and Akt via the G protein–coupled receptors (GPCRs), including GPCR21 (9–12). Growing evidence suggests that these special lipids could function as ligands that activate the GPCRs to elicit a biological function (13, 14). Therefore, if NK cells express these GPCRs, it could be possible that Pls might activate these GPCRs to regulate NK cell functions. To address this, we screened the expression of GPCRs in the human NK cell lines, KHYG-1.

NK cells can kill most abnormal cells within 24 h of their entry into blood circulation. Therefore, NK cells are regarded as the first-line defense in the immune response. CD56^brightCD16^dim NK cells are the major type present in secondary lymphoid tissues, such as the lymph nodes, spleen, Peyer’s patches, and mucosal tissues, where they can be activated by locally produced IL-2 cytokines (15, 16). A fewer number of these cells are circulating in blood compared with potent cytotoxic CD56^dimCD16^bright NK cells, which are mostly present in blood circulation. Also, CD56^brightCD16^dim NK cells present weaker cytotoxic activity before being activated by cytokines such as IL-2. KHYG-1 cells are a type of human CD56^brightCD16^dim NK cell and are known to be a good model to study cytotoxic activity against the lymphoma cells K562 (17–19).

Activated NK cells can induce the expression of cytotoxic molecules such as Perforin-1 and granzymes, resulting in direct killing of virus-infected cells and cancer cells (20, 21). Perforin-1 punches holes in the target cell membrane, allowing passive diffusion of proteolytic molecules called granzymes, which results in apoptosis of the target cells (22, 23). It has been known that cytokine (IL-2) stimulation activates a transcription factor STAT5 in human NK cells. STAT5 is considered one of the key transcription factors that facilitate cytotoxic activity of NK cells by inducing the expression of the perforin-1 (PRF-1) gene (24, 25). IL-2 can induce three different signaling cascades (JAK-STAT, PI3K/Akt/mTOR, and MAPK/ERK) (26), and activation (phosphorylation) of STAT5 is known to be solely regulated by JAKs such as JAK1 and JAK3 (27). JAK activation induces phosphorylation and dimerization of STAT5, which results in nuclear localization and recruitment of STAT5 onto the promoter regions of the target genes, including PRF-1 (28, 29).

In this study, we have investigated the effects of Pls and GPCR21 in the cytolytic function of human NK cells against the target cells and addressed the mechanism of Pls-GPCR21–induced Perforin-1 expression. We have also performed in vivo studies in mice to screen the effect of Pls treatments in host defense against viral invasion and tumor growth.

The KHYG-1 (from JCRB, 0156) and K562 cell lines (from JCRB, 0019) were cultured in RPMI medium with 10% (v/v) FBS (AusGeneX). To enhance activation, we stimulated the KHYG-1 cell line with 10 ng/ml human recombinant IL-2 (R&D Systems). The scallop plasmalogens (sPls) were extracted and purified from scallop as described earlier (30). The sPls are enriched in DHA (20:6) (38.9%) and eicosapentaenoic acid (EPA; 20:5) (23.9%) but also contain ARA (20:4) (9%), oleic acid (OA; 18:1) (2.6%), stearic acid (18:0) (2%), palmitic acid (16:0) (5.2%), and linoleic acids (18:2) (1.2%). The protein kinase A (PKA) inhibitor, H89, was purchased from TOCRIS biotech. Cell proliferation assays were performed by WST-8 assays (Dojindo, Japan). The STAT5 inhibitor (Pimozide) was purchased from R&D Systems.

PBMCs were isolated from two adult heathy volunteers using Lymphoprep by following the written protocol of the provider (STEMCELL Technologies). PBMCs were cultured in RPMI medium with reduced serum (2% FBS) and sPls before treating with IL-2. To see the expression of GPCR21 in the RBC membranes of adult volunteers, we collected blood samples of 10 individuals with reduced blood Pls (average concentration of 1.8 mg/dL) and 10 individuals with normal blood Pls (average concentration of 6 mg/dL). RBCs were isolated by centrifugation and washed three times with the ice-cold PBS buffer (166-23555; Wako). After washing, the cells were hemolyzed with 20 vol of 5 mM phosphate buffer by sonication, and the membranes were collected by centrifugation (13,000 rpm for 10 min). The collected RBC membranes were washed several times by repeated centrifugation to remove hemoglobulin.

Total RNA was extracted from the cells by TRIzol reagent (Life Technologies) according to the recommended protocol. Total RNA was treated with genomic DNA removal kit (Toyobo) and then subjected to cDNA synthesis by PrimeScript 1st strand cDNA synthesis kit (Takara). The PCR amplification of the target genes was carried out by rTaq DNA polymerase (Toyobo) by the following primers: ACTINB, forward 5′-CATCCGCAAAGACCTGTACG-3′ and reverse 5′-CCTGCTTGCTGATCCACATC-3′; IFNA-1, forward 5′-GGAGAGGGTGGGAGAAACTC-3′ and reverse 5′-AAGCGTGACCTGGTGTATGA-3′; IFNA-10, forward 5′-CCAATGGCCCTGTCCTTTTC-3′ and reverse 5′- TGGAGGACAGAGATGGCTTG-3′; IFNG, forward 5′-TGCAGAGCCAAATTGTCTCC-3′ and reverse 5′-TGCTTTGCGTTGGACATTCA-3′; PRF-1, forward 5′-ACCAGGACCAGTACAGCTTC-3′ and reverse 5′-GGGTGCCGTAGTTGGAGATA-3′; GPCR21 (exon 1–exon 2 junction), forward 5′-GCAAGGCTTTCTGCTAAGGA-3′ and reverse 5′-GAGAACGTCTTCAGGCCAAG-3′; GPCR21 (to amplify cDNA encoding GPCR21 protein), forward 5′-CACCTTGGATGGTAATCAGA-3′ and reverse 5′-TCACAATGATGTTGCCAGAA-3′; GPR1, forward 5′-CCCAGGCATGACAATAGCTT-3′ and reverse 5′-TCCCAGAACAAAAGCCAAAC-3′; GPR19, forward 5′-CAAGGGCTATTCCTGACCAA-3′ and reverse 5′-TTGCTCATCCAACTGTGCTC-3′; GPR27, forward 5′-CTACCTGCTGCTCGACCTGT-3′ and reverse 5′-CTCTGCATAGAAGCGGTGGT-3′; and GPR61, forward 5′-GCAGTGCTTCAGGCTTCAGCAG-3′ and reverse 5′-GTTCTGGGCATTACTAACTC-3′.

For mammalian cell expression, human GPCR21 (NM_005294) was cloned by PCR using LA tag polymerase (Takara Bio) with 10% (v/v) DMSO from SH-SY5Y cells cDNA. The PCR-amplified cDNA template was extracted from 1% agarose gel electrophoresis. Simply, the DNA band was cut and put into a 500-μl Eppendorf centrifuge tube (with a hole in the bottom made by needle) filled with a small piece of surgical cotton. The tube with the agarose gel was then placed in a 1.5-ml Eppendorf centrifuge tube and centrifuged at 10,000 rpm for 10 min for the elution of DNA. The eluted DNA samples were then subjected to ethanol precipitation with 3 M Na-acetate. The precipitated DNA was used as a template to clone the open reading frame of GPCR21 by infusion techniques (In-Fusion HD Cloning Kit; Takara Bio) into the CMV promoter–driven expressing plasmids, pcDNA3 (20011; Addgene). Mutations in the N-linked glycosylation sites in GPCR21 extracellular peptides were carried out by site-directed mutagenesis. The mutations were confirmed by DNA sequencing. All primers used to clone wild-type GPCR21 and mutated GPCR21 will be shown on request.

Transfection of the expression plasmids and small interfering RNA (siRNA) was performed by next-generation electroporator CUY21EDIT II (BEX. Co., LTD, Japan). The electroporation parameters were set to five consecutive 10-ms pulses, at 150 V with a 40-ms pause between pulses. Cells (200,000 cells per well of 12-well dishes) were diluted in 100 µl Opti-MEM medium in a cleaned plastic tube and kept on ice before electroporation. The plasmid DNA (1 µg/200,000 cells) and siRNA (10 pmol/200,000 cells) were added and mixed well with the cells by pipetting. After the electroporation, cells were quickly plated in the cultured dish with a warm (37°C) medium. The mission predesigned siRNA targeting GPCR21 (SAS_Hs01_00149534; sense sequence: 5-GTTTCGAATCACTAGTGTA-3′) and GNPAT (SASI_Hs02_00345274; sense sequence: 5-GGCTTATGCTCCAGCACAT-3′) was purchased from Sigma (MERCK).

To detect the expression of endogenous GPCR21, we have screened several lysis buffers and found that the buffer containing sarcosine detergent was the best among them. We used the 0.5% sarcosine lysis buffer to perform all the protein extractions in this study. To disintegrate cells in the lysis buffer, we performed sonication in ice-cold water followed by centrifugation (14,000 rpm for 10 min) to collect dissolved proteins. Protein concentration was measured by bicinchoninic acid protein assay kit (Thermo Scientific), and 20–50 µg proteins of each sample was analyzed by 8–12% NaDodSO₄–PAGE (SDS-PAGE). Proteins were then blotted onto the Clear Trans SP PVDF membranes (FUJIFILM, Japan). The membranes were blocked with TBST buffer supplemented with 5% skim milk for immunoassay (Nacalai, Japan) and then incubated overnight at 4°C with the following Abs: rabbit polyclonal anti-GPCR21 (Invitrogen), mouse monoclonal STAT5 (BD Bio.), mouse monoclonal anti–p-STAT (pY694) (BD Bio.), mouse monoclonal Anti–β-Actin (M177-3; MBL), and rabbit anti–p-ERK1/2 (Cell Signaling, USA). After washing, the membranes were incubated with the secondary Abs (HRP-coupled goat anti-rabbit or anti-mouse IgG) (Cell Signaling) at room temperature for 2 h. The signals from protein were visualized with the Western Lighting Plus-ECL system (PerkinElmer, USA) with LumiCube (Liponics, Japan).

Extraction of phospholipids from the samples and detection of the Pls species, such as the fatty acid components at the sn-1 and sn-2 positions of the glycerol backbone, were conducted by following a previously described protocol (7).

To detect the human Perforin-1 protein in the culture medium, we employed human Perforin ELISA Kit (Diaclone, USA). We followed the recommended protocol to perform ELISA assays.

Lactate dehydrogenase (LDH) is a cytoplasmic enzyme that catalyzes pyruvate to lactate and NADH to NAD⁺. When cells die, LDH is released in the extracellular environments, which can be easily detected by the LDH assays described in the literature (31). In this study, we performed LDH assays with a little modification of the published literature. We first prepared solution A (4 mM p-iodonitrotetrazolium violet [Nacalai] in 0.2 M Tris–HCl [pH 8.2]) and solution B (320 mM Li-lactate [Nacalai] and 6.4 mM NAD [Nacalai] in 0.2 M Tris–HCl [pH 8.2]). The indicator solution (1000×) of 15 mM 1-methoxyphenazine methosulfate (Nacalai) was prepared in 0.2 M Tris–HCl and stocked along with solutions A and B at −20°C. Before LDH assays, solutions A and B were mixed in equal proportion, and the 1-methoxyphenazine methosulfate was added. In each well of a 96-well tissue culture dish, 50 µl of this solution mixture was added to 50 µl of the tested samples (coculture medium of effector, KHYG-1, and target, K562 cells) and incubated for ∼20 min to 1 h in a dark place. Once the color formation was confirmed, 50 µl of the stop solution (1 M acetic acid) was added. The absorbance was measured at 490 nm (A₄₉₀) in the spectrometer (DTX 880 Multimode Detector; Beckman Coulter). Recombinant LDH (Toyobo) was used in each experiment to check the assay performance. The absorbance values were corrected for the background absorbances to quantify LDH activity.

To examine the recruitment of STAT5 proteins onto the promoter regions, we performed chromatin immunoprecipitation (ChIP) assays as described before (7, 32) by using the UPSTATE ChIP kit (Millipore). In each group, 5 × 10⁶ KHYG-1 cells were treated with 1% formaldehyde (final concentration) for 10 min at 37°C. Cells were then subjected to SDS lysis buffer followed by sonication to share DNA to an average length of between 200 and 1000 bp. To pull down the chromatins, we used 5 µl mouse STAT5 Ab (BD Biosciences) and the control normal mouse serum. For the INPUT, we used 10% crossed chromatin samples before the pull-down assays. All the samples were then subjected to the washing steps, followed by elution of the DNA for PCR assays. The primers used for the ChIP assays are as follows: PRF.prom −1150Fw (5′-GTGACAGCTGGAAAGTGATC-3′), PRF1.prom −1001Rv (5′-ATCCCTTGCGTTTATGTCC-3′), PRF1.prom −751Fw (5′-CAGACCACTCTCACCAGCAC-3′), PRF1.prom −601Rv (5′-TCAACCTACATCCCACCCTA-3′), PRF1.prom −571Fw (5′-CATCTCTCTTCTCCCACTCA-3′), PRF1.prom −398Rv (5′-GACTGGTAGGGTTCTTCAGC-3′), GPCR21Fw−234 (5′-CCTGGTGCTATGTGTATGGT-3′), GPCR21Rv−101 (5′-CTGCTGCTCTGGTGCATT-3′), GPCR21Fw−100 (5′-CATCAGGAGCTTGGGGAGTA-3′), and GPCR21Rv+50 (5′-TGGAGTTCATCTTGAGCTCA-3′).

To clone the promoters of human PRF-1 and GPCR21, we used the template genomic DNA extracted from SH-Sy5Y cells. To get the purified genomic DNA, we lysed 1 × 10⁷ SH-SY5Y cells in 500 µl Tail Lysis Buffer (10 mM Tris [pH 8.0], 100 mM NaCl, 10 mM EDTA, and 0.5% SDS) in the presence of Proteinase K (0.5 mg/ml) for overnight at 50°C. After adding 250 µl of 6 M NaCl, the sample was kept on ice for 20 min followed by centrifugation at 10,000 rpm for 10 min. The supernatant was then treated with Isopropanol to get the precipitated genomic DNA. The DNA was washed with 70% EtOH and then dissolved in TE buffer. To amplify the promoter regions, we used 100 ng of purified genomic DNA as a template. LA taq DNA Polymerase (Takara Bio) PCR was used along with 10% DMSO to amplify the promoters successfully. We cloned various fragments of the promoters into the pBV Luciferase vector by the directional cloning using In-Fusion HD Cloning Kits (Takara Bio) as described before (7). All the primers used for the cloning of PRF-1 and GPCR21 promoters will be shown on request. For the promoter assays, we transfected the promoter plasmids into KHYG-1 cells by electroporation. Cells expressing luciferase were lysed with 1× GloLysis buffer (Promega) followed by Luciferase assays with the substrate (Luciferase Assay System; Promega). Luciferase activity was normalized with the protein concentration of the samples. Transfection efficiency of the promoter plasmids electroporation was confirmed each time by cotransfection with EGFP-expressing plasmid.

All animal experiments were approved by the Animal Care Committee of Rheology Institute. For the in vivo tumor growth studies, 12-wk-old male SCID mice that underwent Pls (sPls; 0.02 mg/kg/d) ingestion for 4 wk were injected s.c. in the upper neck with 5 × 10⁶ SH-SY5Y cells suspended with 50% Matrigel matrix (Corning). The oral dose of the sPls was adjusted by considering that each mouse will drink 5 ml of water per day. After 1 mo, mice were sacrificed and examined for tumor weight, histology studies, and blood cell count. For the survival assays, the mice were monitored for 50 d after the tumor cell transplantation.

Mouse CMV (MCMV) was obtained from ATCC (Smith MSGV strain, VR-1399). The virus was propagated by infecting the host SC-1 cell line (RCB1829; Riken BRC cell bank, Japan). For the in vivo virus growth and invasion studies, 6-mo-old male C57BL/6J mice that underwent Pls (sPls; 0.02 mg/kg/d) ingestion for 5 wk were injected i.p. with high-titer MCMV (1 × 10⁸ PFU). After 3 d, mice were sacrificed and examined for blood cell count and virus invasion in the tissues. For the survival assays, mice were monitored for 18 d.

IHC and immunocytochemistry assays were performed according to the previously published protocol (7). Briefly, the mice were transcardially perfused with saline, then with 4% paraformaldehyde, and dissected. The tissues were fixed again with 4% paraformaldehyde overnight, then put into the sucrose solutions (15% followed by 30% w/v sucrose in saline). The submerged tissues were frozen at −80°C with O.C.T. compound (Tissue-Tek). The frozen tissues were sliced at 15 μm by Microtome and placed on glass slides. The slides were stained with Abs against GPR21 (Sigma), CD11b (Abcam), and CD27 (Sigma) by following the protocol (7). To detect the endocytosis of GPCR21, we costained the tissues with Rab5 (early endosome marker; Cell Signaling) and GPCR21 Abs. H&E staining was performed with the frozen tissue sections according to standard protocol. The images were taken by fluorescence microscope (Zeiss Axioskop 2).

To detect the apoptotic cells in the frozen tumor tissue samples, we performed TUNEL assays with In Situ Cell Death Detection Kit, Fluorescein (Roche) by following the standard protocol provided by the manufacturer.

Results were expressed as the mean ± SEM. To examine the p values, we performed one-way ANOVA followed by Bonferroni’s post hoc test for comparison of the paired groups. The p value <0.05 was considered to be statistically significant. For comparison of two groups, we performed Student t test. All the statistical assays, including Kaplan–Meier survival assays, were performed by GraphPad Prism 7 software. The details of the statistics performed in each experiment are mentioned in the figure legends.

GPCR21 is highly expressed in the human NK cell line KHYG-1 (Supplemental Fig. 1A, 1B). Interestingly, GPCR21 is found to be glycosylated in these NK cells compared with the other human cell types (Supplemental Fig. 1C). IL-2 stimulation has been known to induce cytotoxicity of KHYG-1 cells (17). Activated NK cells release cytotoxic molecules such as Perforin-1 and granzymes to kill target aberrant cells (Fig. 1A). The IL-2 stimulation upregulated endogenous expression of cytokines (IFNA-1, IFNA-10, and IFNG), PRF-1, and GPCR21 in KHYG-1 cells (Fig. 1B). IL-2 treatments significantly enhanced N-linked glycosylation of GPCR21 (Fig. 1C–E). N-linked glycosylation was confirmed by tunicamycin treatments (Supplemental Fig. 1D). Knocking down endogenous GPCR21 attenuated IL-2–mediated cytolytic activity of KHYG-1 cells (Fig. 1F–H). In our experimental condition, siRNA transfection reduced endogenous expression of GPCR21 to ∼60% in KHYG-1 cells (Supplemental Fig. 1F). IL-2 treatments enhanced the proliferation of KHYG-1 cells (17), and knockdown of GPCR21 inhibited it (Fig. 1I). The ectopic expression of human GPCR21 enhanced cytolytic activity of KHYG-1 cells against K562 cancer cells (Fig. 1J). GPCR21 overexpression also enhanced the proliferation of KHYG-1 cells (Fig. 1K). These pieces of evidence indicate that GPCR21 can facilitate cytolytic activity of NK cells against the target cancer cells.

Some ethanolamine Pls (PlsEtns) containing fatty acids such as palmitoleic acid (16:0/16:1 PlsEtn), linoleic acid (16:0/18:2 and 18:0/18:2 PlsEtn), OA (16:0/18:1; 18:0/18:1, and 18:1/18:1 PlsEtn), and EPA (16:0/20:5 and 18:1/20:5 PlsEtn) were increased by IL-2 treatments in NK cells (Fig. 2A, 2B). Total PlsEtns were increased by IL-2 treatments in NK cells (Fig. 2C). sPls (Pls derived from scallop) treatments enhanced the proliferation of KHYG-1 cells in the presence of IL-2 (p < 0.05; n = 5) (Fig. 2D). sPls treatments induced cytolytic activity of KHYG-1 cells against the target cancer cells (Fig. 2E). To examine whether sPls treatments, which enhanced cytotoxicity activation and proliferation of NK cells, could change the glycosylation status of GPCR21, we performed immunoblotting assays (Fig. 2F). The extracellular addition of sPls increased glycosylation of GPCR21 (Fig. 2G). In addition, sPls treatments enhanced cytolytic activity of GPCR21-overexpressing KHYG-1 cells (p < 0.01; n = 7) (Fig. 2H). The reduction of endogenous GPCR21 in KHYG-1 cells showed mild attenuation of sPls-mediated cytolytic activity (Fig. 2I). This propounds two possible causes: first, the reduction of GPCR21 by si-GPCR21 (∼60% reduction) might be not enough to prevent the sPls-mediated effects; and second, there might be other receptors that were activated by sPls treatments. However, our present findings suggest that Pls treatments enhance GPCR21-mediated cytolytic activity of NK cells against the target cells.

The ectopic expression of GPCR21 in KHYG-1 cells increased PRF-1 expression (Fig. 3A, 3B). In addition, sPls treatments increased Perforin-1 protein expression in KHYG-1 cells (Fig. 3C). Perforin-1 expression is known to be increased in NK cells on recognition of target cells, such as cancer cells and virus-infected cells. When we cocultured NK cells with cancer cells, Perforin-1 expression was found to be enhanced more in GPCR21-overexpressing KHYG-1 cells than in the control cells (Fig. 3D, 3E). GPCR21-induced cytotoxicity activation was enhanced by sPls treatments in the presence of IL-2 (Fig. 3F), suggesting that Pls-GPCR21 signaling facilitated cytolytic activity of NK cells during the recognition process of the target cancer cells.

sPls treatments enhanced tyrosine phosphorylation (Y694) of STAT5 protein in KHYG-1 cells (Fig. 4A, 4B). STAT5 inhibitor pretreatments attenuated the sPls-mediated induction of PRF-1 expression in KHYG-1 cells (Fig. 4C). To find STAT5 binding sites on PRF-1 genomic promoter regions, we analyzed the University of California Santa Cruz genomic database and identified a cis-regulatory element upstream of the first exon of human PRF-1 gene (Fig. 4D). Notably, a CpG island was identified inside of the second exon (Fig. 4D). However, this study indicated that sPls and IL-2 treatments enhanced PRF-1 gene transcription from the first exon (the primer set was designed in the exon 1–exon 2 junction). We identified three STAT binding consensus sequence regions (TTCnnnGAA, n is for any nucleotide) within 1450 bases upstream of exon 1 (−1072, −654, and −446 position) (Fig. 4D). ChIP assays showed that IL-2 treatments enhanced the recruitment of STAT5 proteins within these consensus regions in human NK cells (Fig. 4E). sPls treatments enhanced this STAT5 recruitment (Fig. 4E), which implies that Pls treatments enhanced STAT5 recruitment onto human PRF-1 promoter regions. Promoter activity of PRF-1 genomic regions was confirmed by luciferase assays in KHYG-1 cells (Fig. 4F). IL-2 treatments enhanced PRF-1 promoter regions, harboring STAT5 binding sites, and sPls treatments enhanced IL-2–mediated activation of PRF-1 promoter activity (Fig. 4F). We then examined whether recognition of the target cancer cells by NK cells could enhance PRF-1 transcription from the promoter region (containing STAT5 binding sites). We overexpressed the PRF-1 promoter-luciferase construct containing two STAT5 binding sites (−693 to first exon) in KHYG-1 cells followed by coculture with K562 cells for 3 h (Fig. 4G). The promoter assays showed a significant increase in transcription from the PRF-1 promoter region in KHYG-1 cells after recognition of the cancer cells (K562) (Fig. 4G). These findings indicate that STAT5-mediated transcriptional activation of the human PRF-1 promoter (upstream of exon 1) in NK cells is tightly regulated by Pls treatments in the presence of IL-2 and by recognition of target cancer cells.

NK cells are known to be activated after physical contact with their target cells, such as cancer cells and virus-infected cells. Target cell recognition enhances activation of NK cells, which facilitates the release of cytolytic molecules, including Perforin-1, to kill the target aberrant cells (Fig. 5A). Glycosylation of GPCR21 was increased within a short period (within 20 min) after physical recognition of target cancer cells by NK cells (Fig. 5B). This evidence suggests that activation of GPCR21 could transform the resting NK cells to the activated state during the recognition process. Interestingly, GPCR21 glycosylation was coupled with phosphorylation of STAT5 and ERK1/2 (Fig. 5B). To identify the N-linked glycosylation site(s) in the GPCR21 peptide sequence, we performed a proteomic study and identified two possible N-linked glycosylation sites in the extracellular domain of GPCR21 (Fig. 5C). We employed point mutations (N/asparagine to R/arginine) by site-directed mutagenesis assays, overexpressed these constructs (wild type and mutants) in KHYG-1 cells, and treated with sPls. Interestingly, we found that sPls treatments did not enhance glycosylation in GPCR21 (8N-R)-expressing KHYG-1 cells, which implies that an amino acid asparagine at position 8 (8N) is involved in glycosylation (Fig. 5C, 5D). In addition, GPCR21 overexpression studies showed that the asparagine mutation (8N-R) in GPCR21 failed to enhance PRF-1 gene expression in KHYG-1 cells (Fig. 5E, 5F), as well as phosphorylation of STAT5 proteins (Fig. 5G, 5H). We then checked cytotoxic activity of KHYG-1 cells, which ectopically expressed the GPCR21 constructs and found that the mutation in the N-linked glycosylation site (8N-R) did not enhance cytolytic activity of NK cells against K562 cancer cells (Fig. 5I). These pieces of evidence imply that glycosylation of GPCR21 is critical in inducing cytotoxicity of human NK cells, and that recognition of target cancer cells possibly acts as a molecular switch to accelerate GPCR21 signaling.

Recognition of K562 cancer cells induced endogenous expression of GPCR21 in KHYG-1 cells (Fig. 6A). We then examined the genomic region (upstream of the first exon) of GPCR21 promoter and identified two STAT binding consensus sequences (Fig. 6B). IL-2 stimulation of NK cells enhanced STAT5 recruitment onto the GPCR21 promoter regions (Fig. 6B). Promoter studies showed that IL-2 stimulation promoted transcriptional activation from the GPCR21 promoter region that had two STAT5 binding regions (Fig. 6C). When we transfected the GPCR21 promoter-luciferase construct in KHYG-1 cells followed by recognition of K562 cells, we found a significant increase in luciferase activity (Fig. 6D). In addition, IL-2–induced GPCR21 gene expression in KHYG-1 cells was attenuated by STAT5 inhibitor treatments (Fig. 6E). These data suggest that activation of human NK cell cytotoxicity can induce transcriptional activation from the human GPCR21 promoter region, which could be because of the increased recruitment of STAT5 onto the GPCR21 promoter.

To investigate whether Pls can regulate IL-2–mediated signaling in human immune cells, we used PBMCs from adult volunteers. Interestingly, sPls pretreatments enhanced IL-2–mediated phosphorylation of STAT5 and ERK1/2 proteins in PBMCs (Fig. 7A). In addition, sPls pretreatments induced PRF-1 expression in PBMCs after treatments with IL-2 (Fig. 7B), which implies that Pls can enhance cytokine sensitivity of human PBMCs to induce PRF-1 expression. Endogenous expression (RNA and proteins) of GPCR21 in human PBMCs was increased by sPls treatments (Fig. 7C, 7D), which suggests that extracellular Pls can regulate endogenous GPCR21 expression in the blood cells. However, we still do not know which cell population (NK cells or other immune cells) in PBMCs is involved in the induction of GPCR21 expression by Pls. Then, we also examined protein expression in the cell membranes of RBCs provided from 18 adult volunteers who were divided into two groups: low-serum Pls group (average 1.8 mg/dl) and normal-serum Pls group (average 6 mg/dl); interestingly, GPCR21 expression was found to be significantly low among the low-serum Pls group compared with the high-serum Pls group (n = 9 per group; p < 0.05) (Fig. 7E, 7F). These data imply that the reduction of blood Pls, which is often associated with age-related diseases, might reduce the expression of GPCR21 in human blood cells, including NK cells.

SCID mice were subjected to in vivo tumor xenograft studies using a human neuroblastoma cell line (SH-SY5Y) to assess possible effects of drinking Pls in limiting tumor growth in vivo. Interestingly, the tumor size and weight were found to be reduced in the Pls group compared with the control group (Fig. 8A, 8B). Drinking Pls significantly reduced the mortality of mice bearing the tumor xenografts (Fig. 8C). In these studies, Pls was given orally by drinking water for 5 wk in a dose of 0.02 mg/kg/d, which is equivalent to 1 mg Pls per day for an adult human weighing 50 kg. A clinical study showed that this dose was effective to enhance the cognitive function of human participants (6). The histology assay showed a disruption of extracellular matrix (ECM) in tumor tissues of the Pls group (Fig. 8D, 8E). CD11b⁺ and CD27⁺ NK cells were increased in tumor tissues of the Pls group (Supplemental Fig. 4). This suggests that drinking Pls might facilitate the entry of NK cells in the tumor regions. It has been known that CD11b is also expressed by macrophages, which indicates that drinking Pls might enhance the recruitment of macrophages in the tumor regions. Interestingly, the IHC data showed that CD11b⁺ and CD27⁺ immune cells expressed GPCR21 and drinking Pls enhanced GPCR21 expression in these immune cells. To further check whether enhanced localization of GPCR21⁺ NK cells is associated with increased apoptosis of tumor cells, we performed TUNEL assays and found a significant increase in apoptosis of tumor cells in the Pls group compared with the control group (Fig. 8F, 8G). The total number of GPCR21⁺ cells was increased in the tumor tissues by Pls treatments (Fig. 8H, 8I), suggesting that oral ingestion of Pls can increase GPCR21 expression in tumor tissues, including in the tumor-resident NK cells, which might enhance cytolytic activity against the tumor cells.

To examine the effects of drinking Pls on viral invasion, we subjected adult C57BL/6J mice to drinking Pls followed by i.p. injection of MCMV (Fig. 9A). The splenic weight increased significantly after MCMV injection, which was attenuated in the Pls ingestion group (Fig. 9B). Drinking Pls rescued the reduction of WBCs in MCMV-infected mice (Fig. 9D). In our experimental condition, there was no change in mouse RBCs and body weight after MCMV infection (Fig. 9C, 9E). Pls ingestion significantly enhanced the survival rate of mice infected with MCMV (data not shown). i.p. injection of MCMV resulted in the spread of viral particles in the lung, splenic, and liver tissues as confirmed by PCR assays using genomic DNA extracted from these tissues (Fig. 9F). The quantity of viral genomic DNA was lower in the Pls group than in the MCMV-infected group, which suggests that drinking Pls may play a role in preventing viral invasion in vivo. We then checked the live viral particles in the tissues by plaque formation assays. Interestingly, drinking Pls significantly reduced the active viral population in the lung, splenic, and liver tissues (Fig. 9G–I). A tissue histology assay showed a reduction of viral clusters in the lung tissues of the Pls group compared with the MCMV group (Fig. 10A, 10B). It is known that MCMV infection is associated with lung inflammation and thickening of the bronchiolar epithelium. In our experiment, we observed a significant increase in thickness of the bronchiolar epithelium in the MCMV group in contrast with the Pls group, which displayed attenuation of this pathological change (Fig. 10A–C). This implies that drinking Pls reduced virus-induced lung inflammation. Similarly, MCMV infection is marked by an increased size of the white pulp in the spleen indicating inflammation in the splenic tissues (Fig. 11A, 11B) and by an increased nuclear size of immune cells in the white pulp (Fig. 11C, 11D). In the red pulp area of the MCMV group, there was an elevated quantity of congested RBCs called extramedullary hematopoiesis, which was reduced in the Pls group (Fig. 11E). To see the mortality rate, we kept monitoring the mice until 20 d and found a significant increase of survival rate in the Pls group mice (Supplemental Fig. 5), suggesting that the Pls treatments could reduce the MCMV-induced toxicity in the experimental mice.

In this article, we report that Pls enhance cytolytic activity of human NK cells by activating GPCR21. To our knowledge, this is the first study to show that GPCR21, especially glycosylated GPCR21, is a potent inducer of cytolytic activity in NK cells. Interestingly, short-time Pls treatments enhanced glycosylation of GPCR21 and increased its endocytosis (data not shown), suggesting that Pls might function as extracellular ligands to activate NK cells via GPCR21. We also noticed an increase in the expression of GPCR21 in NK cells on recognition of cancer cells, which indicates that GPCR21 can be activated when NK cells come in close contact with the target aberrant cells, including cancer cells and virus-infected cells. To gain more insight into how Pls induced Perforin-1 expression in NK cells, we identified transcriptional regulation of the Perforin-1 gene promoter by STAT5 recruitment. The STAT5 recruitment is known to be associated with transcriptional activation of Perforin-1 in different types of cell lines (29). To our knowledge, the involvement of STAT5 in the regulation of Perforin-1 in NK cells had not been known. Our findings were supported by the University of California Santa Cruz genomic database showing cis-regulatory elements at the promoter region of PRF-1. It has been reported that in human PBMCs, Perforin-1 expression is regulated by methylation at the CpG island located in the second exon (33). Although it is still unknown whether Pls can modify genomic methylation, our present findings clearly show that Pls signaling and IL-2–mediated NK cell activation enhance promoter activity from the first exon of the PRF-1 genome, suggesting that Pls might induce PRF-1 transcription without changes in CpG methylation. Further experiments will be necessary to address this issue.

It has been reported that ERK signaling can enhance cytolytic activity of NK cells against the target cells by regulating the expression and release of cytokines (34, 35). In this study, we found that Pls-mediated STAT5 phosphorylation was not canceled by inhibition of PKA signaling in NK cells (Supplemental Fig. 2). In addition, PKA inhibition reduced GPCR21-induced cytotoxicity of NK cells (Supplemental Fig. 2). These pieces of evidence suggest that GPCR21-mediated activation of ERK1/2 signaling, which was dependent on PKA, did not enhance phosphorylation of STAT5 in NK cells on IL-2 stimulation. Interestingly, IL-2 enhanced phosphorylation of ERK, which was independent of PKA signaling. Inhibition of ERK suppressed the ability of NK cells to kill target cancer cells (data not shown), which indicates that ERK signaling can enhance NK cell activity independent of STAT5. These results propose the involvement of two signaling cascades, STAT5 and ERK, as the mechanism of how extracellular Pls could activate cytolytic activity of NK cells against the target cells.

Human blood contains PlsEtns, which are enriched with various long-chain fatty acid–containing Pls (6, 36). To further investigate the functional constituents of Pls species in human blood, which can activate GPCR21 in NK cells, we screened OA-, DHA-, ARA-, and EPA-containing PlsEtns. Interestingly, OA-, ARA-, and EPA-PlsEtns enhanced glycosylation of GPCR21 and induced phosphorylation of STAT5, whereas DHA-PlsEtn failed to do so (Supplemental Fig. 3). This suggests that the fatty acid composition at the sn-2 position of PlsEtns is critical in inducing GPCR21-mediated cytotoxic signaling in NK cells. Lipidomic assays further showed that NK cells released a significant amount of the functionally active Pls species, including OA- and ARA-containing Pls, whereas the target cancer cells did not release Pls (data not shown). In addition, we also found that cytolytic activity of NK cells increased endogenous Pls. These key findings indicate that NK cells can be activated by the Pls-GPCR21 autocrine pathway during recognition of target cells, such as tumor cells and virus-infected cells. So far, there has been presumably no report showing that human NK cells can release functionally active Pls species. This important finding enlightens that NK cells could also be activated via autocrine signaling. Autocrine activation of NK cells by secreted Pls, which can promote the release of Perforin-1, could play a significant role in killing the target cells. Further study will be necessary to address whether NK cells in elderly people have a reduced ability to release the Pls species. Several studies have shown that Pls might function as self-antigen to activate invariant NKT cells (37, 38). However, the Pls-specific receptor had not been identified in these cells. Our present studies could suggest a possibility that invariant NKT cells might express GPCR21, which could regulate the growth and activation of these cells. Further studies will be necessary to address this issue.

In the in vivo experiments, Pls significantly prevented tumor growth and viral invasion with an increased survival rate, suggesting that oral ingestion of Pls might be beneficial to prevent severe symptoms of viral infections including coronavirus disease 2019. Although further experiments will be necessary to address how oral ingestion of Pls boosts immunity, our pieces of evidence of increased expression of GPCR21 in the tumor xenograft of the Pls ingestion group strongly indicate that Pls intake enhances GPCR21 expression in the target tissues, which could lead to apoptosis of tumor cells. To support these findings, we also discovered that GPCR21 was highly expressed in CD11b⁺ and CD27⁺ immune cells including NK cells and macrophages in tumor tissues, and that GPCR21 expression was increased in these cell populations in the Pls ingestion group. This evidence strongly suggests that drinking Pls can enhance the expression of GPCR21 in murine NK cells and macrophages localized in target tumor tissues. The mechanism of how drinking Pls enhanced GPCR21 in murine NK cells remains unknown. Moreover, drinking Pls limited the spread of virus-infected cells in various tissues in mice, implying that drinking Pls might enhance cytolytic activity of murine NK cells to prevent the systemic spread of virus. It has been reported that Pls can be degraded by acidic environments in the stomach and by the enzymatic activity in the cells. Therefore, the mechanism of how drinking Pls reduced the viral spread and tumor growth in these in vivo experiments remains mostly elusive. Our cumulative evidence suggests that Pls can enhance cytolytic activity of NK cells by activating membrane-bound GPCR21 to kill the target aberrant cells and prevent tumor growth and viral invasion in vivo. These findings could indicate that Pls might be beneficial to boost our immune system to fight against viral infection and cancer growth.

We thank Honsho Masanori, Department of Neuroinflammation and Brain Fatigue Science, Kyushu University for providing materials and assisting in DNA sequence analysis. We also thank Aya Satou, Mika Okubo, and Tomomi Morisaki for technical assistance to perform experiments, including liquid chromatography-tandem mass spectrometry in the Institute of Rheology. We thank Yuki Yoneya and Chizuko Kanemaru for their efforts in revising this manuscript.

The authors have no financial conflicts of interest.