Tumor‐derived mitochondrial formyl peptides suppress tumor immunity through modification of the tumor microenvironment

Abstract

Mitochondrial N‐formylpeptides are released from damaged or dead cells to the extracellular spaces and cause inflammatory responses. The role of mitochondrial N‐formylpeptides in aseptic systemic inflammatory response syndromes induced by trauma or cardiac surgery has been well investigated. However, there are no reports regarding the role of mitochondrial N‐formylpeptides in cancer. In this study, we investigated the role of tumor cell‐derived mitochondrial N‐formylpeptides in anti‐tumor immunity using knockout murine tumor cells of mitochondrial methionyl‐tRNA formyltransferase (MTFMT), which catalyze N‐formylation of mitochondrial DNA‐encoded proteins. There was no apparent difference among the wild‐type and MTFMT‐knockout clones of E.G7‐OVA cells with respect to morphology, mitochondrial dynamics, glycolysis and oxidative phosphorylation, oxygen consumption rate, or in vitro cell growth. In contrast, in vivo tumor growth of MTFMT‐knockout cells was slower than that of wild‐type cells. A reduced number of myeloid‐derived suppressor cells and an increase of cytotoxic T‐lymphocytes in the tumor tissues were observed in the MTFMT‐knockout tumors. These results suggested that tumor cell‐derived mitochondrial N‐formylpeptides had a negative role in the host anti‐tumor immunity through modification of the tumor microenvironment.

Tumor cell‐derived mitochondrial N‐formylpeptides had a negative role in the host anti‐tumor immunity through modification of the tumor microenvironment was suggested using knockout tumor cells of mitochondrial methionyl‐tRNA formyltransferase.

1. INTRODUCTION

Recent progress in tumor immunology has expanded the use of immunotherapy, and currently immune checkpoint inhibitor (ICI) therapy is one of the major treatment modalities for cancer. However, the clinical efficacy of the immunotherapy is still insufficient, and only a portion of patients receive clinical benefits. ¹ , ² Thus, further improvement of immunotherapy is an urgent issue.

The tumor microenvironment (TME) influences both the induction and effector phases of tumor‐reactive cytotoxic T‐lymphocytes (CTLs). ³ Most cellular components in the TME, such as regulatory T cells (Tregs), myeloid‐derived suppressor cells (MDSCs), and tumor associated macrophages (TAMs), as well as CTLs, migrate from blood vessels and infiltrate into the tumor tissues. ³ Therefore, tumor cell‐derived chemoattractants and chemokines are important initiators for accumulation of these cellular components in the tumor tissues and subsequent TME formation. Bacterial N‐formylpeptides, such as the formyl‐methionyl‐leucyl‐phenylalanine (fMLF) found in Escherichia coli, are well known chemoattractants for polymorphonuclear cells. ⁴ , ⁵ Formyl peptide receptor (FPR) 1 has been identified as a dominant high‐affinity receptor for the N‐formylpeptides on polymorphonuclear cells. ⁶ FPR1 is a member of the G‐protein coupled receptor family related to signal transduction through intracellular Ca²⁺ flux and subsequent activation of protein kinase C, phosphoinositide 3‐kinase (PI3K), and mitogen‐activated protein kinase (MAPK). ⁶

Mitochondria are thought to have evolutionarily originated from the alphaproteobacterial. ⁷ Protein synthesis in mitochondria as well as in bacteria uses N‐formylmethionine as a starting residue instead of the methionine generally observed in genome‐encoded proteins of eukaryotic cells. ⁸ , ⁹ Thus the mitochondrial N‐formylpeptides have also been shown to exhibit chemoattractant activity in various species, including humans and mice. ¹⁰ , ¹¹ Mitochondrial N‐formylpeptides are categorized as damage associated molecular pattern (DAMP) molecules, also called alarmins, and released from damaged or dead cells to the extracellular spaces. ¹² The role of mitochondrial N‐formylpeptides in aseptic systemic inflammatory response syndromes (SIRS) induced by trauma or cardiac surgery has been investigated by many research groups. ¹³ In contrast, there have been no investigations into the role of mitochondrial N‐formylpeptides in cancer. In regard to FPRs, although there are many reports regarding the roles of FPRs on cancer progression, these studies have only discussed two possible ligands for the FPRs, namely, intestinal microbiome‐derived formylpeptides as pathogen‐associated molecular patterns (PAMPs) and tumor‐derived annexin A1 as DAMP molecules. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷

In the present study, we established murine tumor cells with knockout (KO) of mitochondrial methionyl‐tRNA formyltransferase (MTFMT), which catalyzed N‐formylation of mitochondrial DNA‐encoded proteins, using CRISPR/Cas9 genome editing and investigated the role of tumor cell‐derived mitochondrial N‐formylpeptides in anti‐tumor immunity.

2. MATERIALS AND METHODS

2.1. Mice

Seven‐week‐old female C57BL/6J (B6) and BALB/c‐nu/nu mice were purchased from CLEA Japan (Tokyo) and housed under specific pathogen‐free conditions in the animal facility of Kurume University School of Medicine. All animal experimental protocols were approved using the Institutional Animal Care and Use Committee of Kurume University (approval no. 2019–030) in accordance with the national guidelines on the care and use of laboratory animals. In the tumor transplantation experiments, tumor size was measured every 2 or 3 days. Humane endpoints in this study were as follows: (1) tumor size reached >20 mm in diameter, (2) lethargic condition, and (3) 60 days after tumor transplantation. To obtain tumor tissues or lymphoid tissues, the mice were euthanized by cervical dislocation. The following experiments were conducted using the indicated numbers of mice in each group: in vivo tumor growth analyses using wild‐type (WT) and MTFMT‐KO clones (A1, C9, E3, H10) transplanted into B6 mice (n = 6 per group); tumor growth analyses using WT and two MTFMT‐KO clones transplanted into nu/nu mice (n = 5); immunohistochemical analyses using WT and two MTFMT‐KO clones (n = 3); and FACS analyses of tumor‐infiltrating cells using WT and three MTFMT‐KO clones (n = 4).

2.2. Establishment of MTFMT‐knockout cells

E.G7‐OVA murine lymphoma cells, an EL4 lymphoma‐derived OVA‐expressing cell line, were purchased from the American Type Culture Collection (ATCC) through Summit Pharmaceuticals. The cells were cultured in RPMI 1640 (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% FCS (Thermo Fisher), 0.05 mM 2‐mercaptoethanol, 0.4 mg/mL G418, and 50 μg/mL gentamicin at 37°C in a 5% CO₂ incubator. Knockout of the MTFMT gene in E.G7‐OVA cells was performed using an MTFMT‐KN2.0 mouse gene knockout CRISPR kit (KN510471; Origene, Rockville, MD) according to the manufacturer's instructions. The 2 × 10⁶ cells were transfected with 2 μg of pCas‐Guide CRISPR vector containing MTFMT guide RNA (gRNA) and 5.5 μg of linear donor EF1a‐GFP‐P2A‐Puro by electroporation using a Neon Transfection System (Thermo Fisher) under the following parameters: pulse number 1, 1,080 V, 50 ms, 100‐μL tip, and 0.1 mL Buffer R. The target sequences of the gRNA vectors supplied in the kit were 5′‐AAGCACCTGCCGGGTTCGAC‐3′ for vector 1, and 5′‐ACGTCGGGCCACTCGTACAC‐3′ for vector 2; these sequences targeted mRNA positions 65–84 in exon 1 and positions 304–323 in exon 2, respectively. The linear donor contains a stop codon, and thus the insertion of the linear donor at the editing site in exons 1 and 2 disrupts all the functional domains of MTFMT. Puromycin selections (10 μg/mL) of the cells were performed at 6, 9 or 13 days after the transfection. Single‐cell colonies were obtained by limiting dilution after the puromycin selection. Knockout of the MTFMT gene was confirmed using western blot analysis. The knockout clones were maintained in the presence of puromycin (10 μg/mL) and used for further experiments before 1 month of in vitro culture. MTFMT‐knockout clones were also established from B16F10 murine melanoma cells (ATCC) using AAVpro CRISPR/SaCas9 helper free system (Takara Bio, Kusatsu, Japan) according to the manufacturer's instruction. The sequences of the gRNA were designed using CRISPOR guide RNA design tool (http://crispor.tefor.net/crispor.py): 5′‐CTGTTCCTGTCAAGCACCTG‐3′ and 5′‐GGGAAGGACGGCCGAGGTTC‐3′; these sequence targeted mRNA positions 54–73 and 85–104 in exon 1, respectively. All gene modification experimental protocols were approved by the Institutional Genetic Modification Safety Committee of Kurume University (approval no. 31–16) in accordance with the national guidelines for research involving recombinant DNA experiments.

2.3. Western blotting

Western blotting was performed as previously described. ¹⁸ In brief, cells were lysed using RIPA buffer (Thermo Fisher) with 1× protease inhibitor cocktail (Nacalai Tesque) for 5 min on ice, then sonicated and centrifuged at 14,000  g for 15 min at 4°C to remove the cell pellet. The protein contents of the lysate were determined using a BCA protein assay kit (Thermo Fisher), and 8 μg of protein for each lysate was mixed with 4× NuPAGE LDS Sample Buffer (Thermo Fisher) and 10× NuPAGE Reducing Agent containing 0.5 M dithiothreitol, denatured at 70°C for 10 min, and subjected to 12% SDS‐PAGE. After electrophoresis, the proteins were transferred to an Immobilon‐P membrane (Merk Millipore, Darmstadt, Germany). The membrane was blocked with Blocking One (Nacalai Tesque) at room temperature (r.t.) for 1 h and then incubated with 1:700 rabbit anti‐MTFMT antibody (364,231; USBiological, Salem, MA). After washing with 0.1% Tween 20‐Tris‐buffered saline (TBST), the membrane was incubated with 1:5000 horseradish peroxidase‐conjugated anti‐rabbit IgG (ab6721; Abcam) for 2 h at r.t. and rinsed with TBST three times. Detection was performed using Clarity Western ECL Substrate (Bio‐Rad, Hercules, CA) and a LAS‐4000 mini system (Fujifilm). The band intensities were analyzed using MultiGauge v. 3.0 (Fujifilm). The expression of MTFMT protein was normalized to the amount of loading protein or β‐actin using anti‐β‐actin (1:5000; ab8227, Abcam, Cambridge, UK). COS‐7 cells were transfected with mouse MTFMT cDNA clone (EX‐Mn34948‐Lv207; GeneCopoeia, Rockville, MD) according to the manufacturer's instructions and the cell lysate was used as a reference.

2.4. Assessment of in vitro cell growth, glycolysis, oxidative phosphorylation, oxygen consumption rate, and reactive oxygen species

MTFMT‐KO or WT cells were placed into the wells of a 96‐well plate (1 × 10⁴ cells/well), and the live cell number in the culture was counted daily using a cell counting kit‐8 (Dojindo, Kumamoto). The glycolysis and oxidative phosphorylation (OXPHOS) of the cells were determined using a Glycolysis/OXPHOS Assay Kit (Dojindo) according to the manufacturer's instructions. In brief, cells in 100 μL of 10% FCS‐RPMI1640 medium were placed into the wells of a 96‐well plate (2 × 10⁴ cells/well) and incubated at 37°C for 5 h with or without 1.25 μM oligomycin. After incubation, the cell culture supernatants were subjected to lactate assay. An assay mixture containing nicotinamide adenine dinucleotide (NAD), lactate dehydrogenase (LDH), and WST, a water soluble tetrazolium salt, was added to the supernatant and incubated at 37°C for 30 min, and absorbance at 450 nm was measured using a plate reader (1420 Multilabel counter ARVO MX; PerkinElmer, Shelton, CT). For ATP assay, 1 × 10⁴ cells of the culture were transferred to a well of a white 96‐well plate and a working solution containing luciferase and luciferin was added to the wells. After 10 min incubation at 25°C, luminescence was measured using a luminometer (GloMax‐Multi+ Detection System with Instinct Software ver.3.0; Promega, Madison, WI). Data from triplicate assays are shown. The oxygen consumption rate (OCR) of the cells was determined using an Extracellular OCR Plate Assay Kit (Dojindo) according to the manufacturer's instructions and the phosphorescence intensity was measured using an Infinite 200 Pro M Nano (TECAN) fluorescence microplate reader. The maximum and minimum respiratory capacities of the cells were assessed by the addition of 2 μM carbonyl cyanide 4‐(trifluoromethoxy)phenylhydrazone (FCCP) and 10 μM antimycin A, respectively. The reactive oxygen species (ROS) of the cells were determined using a dihydroethidium (DHE) Assay Kit – Reactive Oxygen Species (Abcam) according to the manufacturer's instructions. Cells were labeled with 5 μM DHE, and 10 μM antimycin A and 20 mM N‐acetyl cysteine were used as positive and negative controls, respectively. The ROS were evaluated using flow cytometry.

2.5. Assessment of in vivo tumor growth

Next, 1 × 10⁶ WT or MTFMT‐KO cells were subcutaneously (s.c.) injected into the flanks of mice, and tumor size was measured every 2 or 3 days using a caliper. To assess the effect of formyl peptides on in vivo tumor growth, 1 × 10⁶ WT or 1 × 10⁷ MTFMT‐KO cells were s.c. injected into the flanks of mice, and a mixture of murine mitochondrial formyl peptides (mMT‐FPs) were daily injected to the intratumor (i.t.) after day 4 or 5 of the tumor transplantation. The mMT‐FP mixture consisted of 50 μg murine mitochondrial formyl peptide ND1 (formyl MFFINI) and 50 μg ND5 (formyl MINIFTT) in a volume of 0.1 mL. These custom peptides were purchased from GenScript (Tokyo).

Tumor volumes were calculated using the following formula:

$$\text{Tumor volume}\left({\mathrm{mm}}^3\right)=\left(\text{greatest longitudinal diameter}\right)\times {\left(\text{greatest transverse diameter}\right)}^2\times 0.5.$$

Two independent experiments were performed using 5–7 mice per group.

2.6. Immunohistochemistry

Immunohistochemistry was performed as previously described. ¹⁸ In brief, tissue specimens were fixed in 10% neutral buffered formalin, paraffin‐embedded, and cut into 4‐μm sections. The sections were subsequently transferred to glass slides, deparaffinized, and rehydrated. Antigen retrieval was performed in 1 mM EDTA–10 mM Tris–HCl (pH 9.0) at 110°C for 30 min. After blocking of endogenous peroxidase with BLOXALL (Vector Laboratories, Burlingame, CA) followed by 2.5% normal horse or goat serum for 20 min, the sections were further incubated with the first antibodies at r.t. for 1 h or at 4°C overnight, rinsed twice with 0.1% Tween 20–phosphate‐buffered saline for 5 min, and incubated with the secondary antibodies at r.t. for 30 min. Chromogenic detection was performed using HistoGreen (HISTOPRIME, Linaris Biologische, Mannheim, Germany). Slides were counterstained with Vector Hematoxylin QS (Vector Laboratories). The antibodies used for the immunostaining were as follows: rabbit mAbs against mouse CD4 (1:500; EPR19514; Abcam), CD8α (1:500; EPR20305; Abcam), F4/80 (1:350; D2S9R; Cell Signaling Technology, Tokyo), CD11c (1:200; D1V9Y; Cell Signaling Technology), and rat anti‐Ly‐6G/Ly‐6C (1:200; RB6‐8C5; Novus Biologicals, Centennial, CO). Peroxidase‐conjugated horse anti‐rabbit IgG and anti‐rat IgG polymer kits (ImmPRESS; MP‐7801 and MP‐7444; Vector Laboratories) were used as the secondary antibodies. For the staining of CD11c, anti‐CD11c mAb was incubated at 4°C overnight, and Boost IHC detection reagent (HRP, rabbit; Cell Signaling Technology) was used as the secondary antibody for 30 min at r.t., following the 15 min incubation at r.t. with EnVision FLEX+ rabbit LINKER (Agilent, Santa Clara, CA).

2.7. Mitochondrial imaging

Cells were incubated in 10% FCS–RPMI 1640 containing 0.1 μM MitoBright LT red for microscopic analysis or deep red for flow cytometric analysis (Dojindo, Kumamoto, Japan) for 30 min at 37°C. The cells were then washed twice and analyzed using a laser confocal microscope (FV1000; Olympus, Tokyo) or a flow cytometer (see below).

2.8. Flow cytometric analysis

Tumor‐infiltrating cells (TILs) were prepared as follows. Tumor tissue specimens were minced and incubated in 2.38 mL RPMI 1640 with 100 μL of the Enzyme D, 10 μL of the Enzyme R, and 12.5 μL of the Enzyme A supplied in a Tumor Dissociation Kit, mouse (Miltenyi Biotec, Bergisch Gladbach, Germany) at 37°C for 30–40 min. After removing cell debris using a 70‐μm strainer, cells were washed and suspended in 10% FCS–RPMI1640. Mononuclear cells were further separated by density gradient centrifugation using Lympholyte‐M (Cedarlane, Burlington, NC). The cells were incubated for 30 min on ice with the appropriate dilution of antibodies. For the analysis of Foxp3 expression, the cells were stained with anti‐CD4 and anti‐CD8, then treated with True‐Nuclear™ Transcription Factor Buffer Set (BioLegend, San Diego, CA) and further stained with anti‐Foxp3 antibody according to the manufacturer's instruction. The antibodies used in this study were as follows: anti‐CD4‐FITC (clone GK1.5), anti‐CD8‐PerCP‐Cy5.5 (clone 53–6.7), anti‐CD25‐APC/Cy7 (clone PC61), anti‐Foxp3‐PE (clone MF‐14), anti‐CD11b‐Alexa Fluor488 (clone M1/70), anti‐CD11c‐PerCP/Cy5.5 (clone N418), anti‐I‐A/I‐E‐PE (clone M5/114.15.2), anti‐Ly‐6C‐PerCP/Cy5.5 (clone HK1.4), anti‐Ly‐6G‐APC/Cy7 (clone 1A8), anti‐F4/80‐APC/Cy7 (clone BM8), anti‐CD206‐PE (clone C068C2), and anti‐CD279 (PD‐1)‐APC/Cy7 (clone 29F.1A12) from BioLegend. The isotype control antibodies were as follows: APC/Cy7 rat IgG2a (clone RTK2758), PE rat IgG2b (clone RTK4530), PE rat IgG2a (clone RTK2758), PerCP/Cy5.5 Armenian hamster IgG (clone HTK888), and PerCP/Cy5.5 rat IgG2c (clone RTK4174) from BioLegend. The stained cells were analyzed on a BD FACS Canto II flow cytometer with FACS Diva software (BD Biosciences). For the analysis of Ki67 expression in the tumor cells, the cells were treated with Foxp3/transcription factor staining buffer set (Invitrogen) and further stained with anti‐Ki67‐FITC (clone SoIA15, Invitrogen).

2.9. Assessment of CTLs

Cells from the inguinal lymph nodes or TILs at 9 days after s.c. tumor transplantation were subjected to CTL assessment. 5 × 10⁵ lymph node cells or 1.3 × 10⁵ TILs in 0.2 mL of X‐vivo 15 (Lonza, Basel, Switzerland) supplemented with 2 mM l‐glutamine, 10 mM HEPES and 50 μg/mL gentamicin were cultured in the presence or absence of 10 μg/mL OVA peptide (SIINFEKL) at 37°C for 18 h in a well of a 96‐well MutliScreen HA plate (Millipore, Billerica, MA) coated with 10 μg/mL of anti‐mouse IFN‐γ mAb (AN18, Mabtech, Nacka Strand, Sweden). After washing away the cells, the plate was incubated with 1 μg/mL of biotinylated anti‐mouse IFN‐γ Ab (R4‐6A2, Mabtech) for 2 h at r.t., followed by 1 h incubation of 1:1000‐diluted ExtrAvidin‐alkaline phosphatase (Sigma‐Aldrich, St. Louis, MO) at r.t. The IFN‐γ‐producing cell spots were visualized using a SIGMAFAST BCIP/NBT tablet (Sigma‐Aldrich) and counted using an ImmunoSpot analyzer (Cellular Technology, Cleveland, OH). In some experiments, CD8⁺ T cells were isolated from the lymph node or TILs using a CD8a⁺ T Cell Isolation Kit, mouse (Miltenyi Biotec) and subjected to CTL assessment.

2.10. In vivo depletion of T cells

Mice were intraperitoneally injected in total three times with 0.25 mg/mouse of anti‐CD3 (clone 2C11) monoclonal antibody (mAb) from Bio X Cell (Lebanon, NH), on the day of tumor inoculation (day 0) and days 4 and 8. Depletions of the CD3 cells using this protocol were confirmed using flow cytometric analysis of spleen cells obtained 1 day after the mAb inoculation (data not shown). Two independent experiments with five mice per group were performed.

2.11. Statistical analysis

Differences between the two groups were analyzed as follows. First, the data shown in Figure 2 were analyzed using two‐way repeated measures ANOVA. When the interaction between groups and points in time was statistically significant, linear regression analyses were further performed to assess the slope of each subject, and the averages of the slopes in each group were compared using the Dunnett test (Figure 2A) or Student's t‐test (Figure 2B,C for C9 and H10). The data in Figure 3 were analyzed using the Dunnett test. A p‐value < 0.05 was considered statistically significant. Statistical analyses were performed using JMP Pro version 16 software (SAS Institute, Cary, NC).

FIGURE 2.

Tumor growth of in vivo transplanted cells of WT and MTFMT‐KO clones. Representative results of at least two experiments are shown. (A) Tumor growth of WT and MTFMT‐KO clones (A1, C9, E3, H10) after s.c. transplantation to B6 mice is shown (n = 6 per group). (B) A mixture of murine mitochondrial formylpeptide (mMT‐FPs) or vehicle control (0.2% DMSO) was daily injected i.t. beginning on day 4 or 5 after transplantation of WT or H10 cells into B6 mice (n = 6). 1 × 10⁶ cells of WT (left panel) and H10 (center panel) and 10 × 10⁶ cells of H10 (right panel) were s.c. transplanted. (C) Tumor growth of WT, C9 and H10 cells in B6, athymic nu/nu and CD3‐depleted mice (n = 4 or 5). Representative results of at least two experiments are shown. Error bars represent the standard error of the mean. (D) Reactive oxygen species (ROS) of the WT and H10 cells from in vitro culture and from tumor‐bearing mice were measured. Antimycin A and N‐acetyl cysteine (with antimycin A) were used as positive and negative controls, respectively. The left panel shows representative histograms of in vitro cultured WT cells. (E) Effect of i.t. injection of mMT‐FPs on the tumor growth of WT and H10 cells in nu/nu mice. n.s., not significant.

FIGURE 3.

Induction of CTLs in the WT and MTFMT‐KO tumor‐bearing mice. Lymph node cells or tumor‐infiltrated lymphocytes (TILs) from WT or MTFMT‐KO tumor‐bearing mice at day 9 after tumor cell transplantation were subjected to IFN‐γ ELISPOT assay. (A) Number of IFN‐γ‐producing SPOTs in the lymph nodes. Unseparated (left panel) or CD8⁺ purified (right panel) lymph node cells were used for the assay. (B) Number of IFN‐γ‐producing SPOTs in the TILs. (C) PD‐1 expression on CD8⁺ cells of the lymph node from WT and H10 tumor‐bearing mice. Lymph nodes from four mice were pooled and used for analysis. (D) Number of IFN‐γ‐producing SPOTs in the TILs after normalization by the tumor weight. Tumor tissues from four mice were pooled and used for the preparation of TILs. The weights of the pooled tumor tissues are also shown. n.s., not significant.

3. RESULTS

3.1. Establishment of MTFMT‐knockout clones

The MTFMT gene of E.G7‐OVA cells was disrupted using a CRISPR/Cas9 system with pCas‐Guide vectors. We used two gRNA sequences (gRNA 1 and gRNA 2) that were respectively located at mRNA positions 65–84 of exon 1 and 304–323 of exon 2. The protein expression of MTFMT in the KO clones was further confirmed using western blot analysis (Figure 1A). The 37 KDa MTFMT bands were detected in both WT E.G7‐OVA cells and MTFMT gene transfected COS‐7 cells, but not in the MTFMT‐KO clones.

FIGURE 1.

Establishment of MTFMT‐knockout clones of E.G7‐OVA cells. (A) Western blot analysis of MTFMT‐knockout clones of E.G7‐OVA cells. A 37 kDa band of MTFMT protein found in the wild‐type (WT) E.G7‐OVA cells and MTFM‐transfected COS‐7 cells was not detected in knockout clones. β‐Actin was used as an internal control. Experiments for each clone were repeated at least twice during screening and representative results are shown. (B) Representative images under phase‐contrast microscope observation of WT and MTFMT‐KO clones of E.G7‐OVA cells. (C) Mitochondria were stained with MitoBright LT red for laser confocal microscopy or deep red for flow cytometry and representative images and histograms are shown. (D) Amounts of the cellular ATP and lactate in the culture supernatants of the WT and MTFMT‐KO clone H10 cells were measured. Cellular ATP levels were normalized against the levels of no‐inhibitor controls and shown as a relative cellular ATP amount. Oligomycin was used as an OXPHOS inhibitor. (E) Oxygen consumption rates (OCR) of the WT and H10 cells were measured. FCCP and antimycin were used as a protonophore and an inhibitor of mitochondrial electron transport. The right panel shows relative OCR after normalization against the levels of no‐inhibitor controls. (F) In vitro cell growth and Ki67 staining of WT and MTFMT‐KO clones of E.G7‐OVA cells. There was no significant difference in cell growth and proliferation among the WT and MTFMT‐KO clones. Representative results are shown. n.s., not significant.

3.2. Characteristics of MTFMT‐KO clones

Morphological changes were not detected under phase‐contrast microscope observation between WT and MTFMT‐KO cells. Representative images are shown in Figure 1B. The effect of MTFMT‐KO on mitochondrial dynamics was further analyzed. Mitochondria of WT and MTFMT‐KO cells were stained using MitoBright LT dye and analyzed using laser confocal microscopy and flow cytometry. As shown in Figure 1C, there were no differences between WT cells and MTFMT‐KO clones in either analysis, suggesting that MTFMT‐KO did not affect the mitochondrial dynamics. We also investigated the effects of MTFMT‐KO on the glycolysis and oxidative phosphorylation (OXPHOS) of the WT and MTFMT‐KO cells. The relative contribution of mitochondrial OXPHOS to the total ATP synthesis estimated by the addition of oligomycin, a mitochondrial OXPHOS inhibitor, of the WT and MTFMT‐KO clone H10 cells were respectively 7.7% and 8.9%, and the difference was not statistically significant (Figure 1D). Compensatory glycolysis under suppression of mitochondrial OXPHOS by oligomycin was further analyzed by measuring the lactate production. Lactate levels in the culture supernatants of the WT and clone H10 cells were increased by 32.7% and 43.0%, and the difference was not statistically significant (Figure 1D). The OCRs of the WT and KO cells were also measured, and again there was no significant difference between the WT and KO cells (Figure 1E). These results suggest that MTFMT‐KO does not affect the mitochondrial functions. There were also no significant differences in in vitro cell growth or Ki67 expression that correlated with cell proliferation among the WT cells and MTFMT‐KO clones (Figure 1F).

3.3. In vivo tumor growth of WT and MTFMT‐KO cells and characteristics of tumor‐infiltrated cells

The in vivo tumor sizes of subcutaneously transplanted WT or MTFMT‐KO cells in B6 mice were measured every 2 or 3 days. The in vivo tumor growth of the MTFMT‐KO clone cells was markedly suppressed when compared with that of WT cells (Figure 2A). This suppression was observed in both the KO clones using gRNA1 and gRNA2 for editing. Similar tumor growth suppression by MTFMT‐KO was observed in B16F10 melanoma cells produced using a different gRNA/CRISPR‐Cas9 system (Figure S1). These facts suggest that the tumor growth suppression observed in the MTFMT‐KO clones was not due to an off‐target effect of gene editing. To clarify the reason for the difference in results between the in vitro cell growth and in vivo tumor growth of MTFMT‐KO cells, we investigated the effect of co‐culture of tumor cells with normal spleen cells (1:1 to 1:5), mimicking the tumor microenvironment, on the in vitro cell growth and viability of the tumor cells. Neither the cell growth nor the viability of the WT and MTFMT‐KO cells was not affected by the co‐culture with normal spleen cells (data not shown). To determine whether the tumor growth suppression observed in MTFMT‐KO clones was due to a depletion of mitochondrial N‐formylpeptides, a mixture of murine mitochondrial formyl peptides (mMT‐FPs) was i.t. injected every day after day 4 or 5 of tumor transplantation, and the tumor growth was recorded (Figure 2B). The left‐hand two panels show the results when 1 million WT or MTFMT‐KO clone H10 cells were s.c. transplanted to the mice. In both the WT and H10 groups, no effect of mMT‐FPs i.t. injection was observed. The deviation of tumor sizes of the H10 group was relatively large since in this group the tumors were too small or were almost rejected. Therefore, 10 × 10⁶ H10 cells were s.c. transplanted and the effect of mMT‐FPs i.t. injection was re‐examined (right panel). As expected, the deviation of tumor sizes declined to an acceptable level. Under this condition, the tumor growth in the H10 group appeared to be enhanced by the i.t. injection of mMT‐FPs compared with that of the DMSO controls (right panel). These results suggested that the tumor growth suppression observed in the MTFMT‐KO clones was due to the depletion of mitochondrial N‐formylpeptides in the MTFMT‐KO cells.

The suppression of the tumor growth observed in MTFMT‐KO clones in B6 mice was recovered in nu/nu mice and CD3‐depleted mice (Figure 2C), and thus the suppression was mainly mediated by host T‐cell‐mediated immunity. However, it remained a possibility of the influence of MTFMT‐KO on the mitochondrial functions in MTFMT‐KO cells and it affected the in vivo tumor growth. We analyzed the basal levels of ROS as a mitochondrial function of tumor cells from tumor‐bearing mice and in vitro cultured cells using flow cytometry (Figure 2D). There was no apparent difference in the ROS activities between WT and H10 cells from tumor‐bearing mice, and a similar trend was observed in the in vitro cultured cells. These results suggest that the other mitochondrial functions of in vitro cultured MTFMT‐KO cells shown in Figure 1D,E may be similar to those of the cells from tumor‐bearing mice.

The effect of i.t. injection of mMT‐FPs on the tumor growth of WT and H10 cells in nu/nu mice was further examined (Figure 2E). In nu/nu mice, i.t. injection of mMT‐FPs had no apparent enhancement effect on the tumor growth of WT and MTFMT‐KO clones. These results suggest that mMT‐FPs do not affect the characteristics of tumors.

3.4. Induction of CTLs in the WT and MTFMT‐KO tumor‐bearing mice

Cells of the lymph nodes and TILs were obtained from WT or MTFMT‐KO E.G7 tumor‐bearing mice at day 9 of tumor cell transplantation and subjected to IFN‐γ ELISPOT assay in the presence of an H‐2K^b‐restricted OVA peptide (SIINFEKL) (Figure 3). The numbers of IFN‐γ‐producing SPOTs in the lymph nodes of MTFMT‐KO tumor‐bearing mice were lower than those in the WT tumor‐bearing mice (Figure 3A, left panel). When the purified CD8⁺ cells of lymph nodes from WT and H10 tumor‐bearing mice were subjected to the ELISPOT assay, a similar trend was observed (Figure 3A, right panel). Namely, the IFN‐γ‐producing SPOT numbers in the TILs of MTFMT‐KO tumor‐bearing mice were similarly lower than those in WT tumor‐bearing mice (Figure 3B). One of the possible reasons for the lower IFN‐γ‐producing SPOT numbers in the MTFMT‐KO tumor‐bearing mice may be the exhaustion of T cells. As expected, the level of PD‐1 expression on the CD8⁺ lymph node cells from H10 tumor‐bearing mice was higher than that in the WT tumor‐bearing mice (3.9% and 2.0%, respectively; Figure 3C). As shown in Figure 2A, the tumor sizes of MTFMT‐KO clones were markedly smaller than those of the WT tumors, and therefore the IFN‐γ‐producing SPOT numbers in the TILs were further adjusted for tumor weight (Figure 3D). This compensation altered the relation of the number of SPOTs in the TILs of MTFMT‐KO clones to that of WT tumors, with the result that the SPOT numbers in the TILs of MTFMT‐KO tumor‐bearing mice became higher than those of WT tumor‐bearing mice.

3.5. Immunohistochemical and flow cytometric analyses of tumor‐infiltrated cells

Tumor specimens of WT and MTFMT‐KO clones obtained from B6 mice were further analyzed using immunohistochemistry (Figure 4). Infiltration of F4/80⁺ macrophages and CD11c⁺ cells, mainly containing dendritic cells, was markedly increased in MTFMT‐KO tumors. Increases of CD4⁺ T cells, CD8⁺ T cells and Gr1⁺ granulocytes were also observed in MTFMT‐KO tumors. The cell counts per area of each population are shown in Figure 4B. In addition to IHC, we further conducted a flow cytometric analysis of TILs of WT and H10 tumors. Tumor tissues from four mice were pooled and subjected to the flow cytometric analysis and the results are shown in Figure 4C. The percentages of each population in the TILs are shown in the left panel. Relative contents of F4/80, CD11c and Gr1 in the TILs of H10 appear to be lower than those of the WT tumor. We thought the discrepancy between the results of IHC and flow cytometric analysis was due to the different yields of TILs of WT and H10 tumors. The cell number of each population in the tumor tissues was further normalized against tumor weight (Figure 4C, right panel). The adjusted results shown in this panel also indicated that the CD4, CD8, F4/80, CD11c, and Gr1 populations were increased in the H10 tumor tissues. CD11b⁺ subsets and regulatory T cells (Tregs) in the tumor tissues were subsequently analyzed using flow cytometry. Representative staining patterns and percentage of each population against the total number of TILs are shown in Figure 5A,B, respectively. CD11b⁺ cells were divided into F4/80 ⁺ macrophages and Gr‐1⁺ MDSCs. Because the anti‐Gr‐1 antibody recognizes both Ly‐6C and Ly‐6G, we used anti‐Ly6C and ‐Ly6G mAbs. CD11b⁺ F4/80⁺ macrophages were subdivided into CD11c⁺ type‐1 macrophage (M1) and CD206⁺ type‐2 macrophage (M2) subsets, and CD11b⁺ MDSCs were subdivided into Ly‐6C^hi Ly‐6G⁻ monocytic MDSCs (M‐DSCs) and Ly‐6C^lo Ly‐6G⁺ polymorphonuclear MDSCs (PMN‐MDSCs). A dominant subset of macrophages was M2 in both the WT and MTFMT‐KO tumor tissues (1.8 ± 0.18% and 1.3 ± 0.12%, respectively) and only a small portion of total TILs were M1 (0.3 ± 0.08% and 0.3 ± 0.04%, respectively). Thus, there were no differences in tumor‐infiltrated macrophages between the WT and MTFMT‐KO tumor tissues. The dominant MDSCs were PMN‐MDSCs in both the WT and MTFMT‐KO tumor tissues, and the content of PMN‐MDSCs in MTFMT‐KO tumor tissue (1.7 ± 0.33%) was significantly lower than the level in WT tumor tissue (10.3 ± 0.81%, p = 0.003). There were no significant differences in the contents of M‐MDSCs and Tregs between the WT and MTFMT‐KO tumor tissues.

FIGURE 4.

Tumor specimens at day 9 after tumor cell transplantation were subjected to immunohistochemistry and flow cytometric analysis. (A) Representative images of the immunohistochemistry of WT and H10 tumor tissues are shown. Scale bars in the upper and lower panels are 500 and 200 μm, respectively. (B) Counts of positively stained cells per area. (C) Tumor tissues from four mice were pooled and subjected to flow cytometric analysis. The left and right panels show % TILs and number of cells per tumor tissue weight, respectively. The numbers of TILs per 1 g of WT and H10 tumor tissue were respectively 0.46 × 10⁶ and 4.92 × 10⁶ cells in this experiment. n.s., not significant.

FIGURE 5.

Flow cytometric analyses of CD11b⁺ subsets and regulatory T cells (Tregs) in the tumor tissues. CD11b⁺ cells were divided into F4/80 ⁺ macrophages and Gr‐1⁺ (Ly‐6C⁺ and/or Ly‐6G⁺) myeloid‐derived suppressor cells (MDSCs). CD11b⁺ F4/80⁺ macrophages were subdivided into CD11c⁺ type‐1 macrophage (M1) and CD206⁺ type‐2 macrophage (M2) subsets, and CD11b⁺ MDSCs were subdivided into Ly‐6C^hi Ly‐6G⁻ monocytic MDSCs (M‐MDSCs) and Ly‐6C^lo Ly‐6G⁺ polymorphonuclear MDSCs (PMN‐MDSCs). (A) Representative staining profiles of TILs from WT and H10 tumor tissues. A trend similar to that in the other MTFMT‐KO clones was observed. (B) Percentage of each population against the total number of TILs. (C) The mMT‐FPs or vehicle control were daily injected i.t. beginning on day 4 after transplantation of WT or H10 cells into B6 mice. Tumor tissues were subjected to flow cytometric analysis (n = 3). n.s., not significant; P‐MDSC, PMN‐MDSCs.

The effect of i.t. injection of mMT‐FPs on the TIL populations of WT and H10 tumors was analyzed using flow cytometry (Figure 5C). An increase of PMN‐MDSCs was observed in WT (p = 0.023) in the mMT‐FPs injected groups. A similar tendency, but not statistically significant (p = 0.055), was also observed in H10 tumors. With respect to the tumor‐infiltrated macrophages, M2 macrophages were increased in WT tumors after mMT‐FPs injection (p = 0.044) and a similar tendency was observed in H10 tumors. A slight increase of M1 cells was also observed in H10 tumors after mMT‐FPs injection (p = 0.028). An increase of Tregs was not observed in both WT and H10 tumors in the mMT‐FPs injected group.

4. DISCUSSION

HMGB1 is one of the major chromatin‐associated non‐histone proteins in the nucleus, ¹⁹ and several receptors for HMGB1 have been identified, i.e., the Toll‐like receptors (TLR)‐2 and TLR−4, the receptor for advanced glycation end products (RAGE), and T‐cell immunoglobulin and mucin‐domain containing‐3 (TIM‐3). ¹⁹ Knockout of HMGB1 in the tumor cells has been shown to suppress in vivo tumor growth through host CTL‐mediated immunity by enhancing the infiltration of immune cells into the tumor tissues, thereby converting tumors from “cold” tumors, in which immune‐related cells are poorly infiltrated, to “immune‐inflamed” or “hot” tumors. ¹⁸ These results suggested that tumor cell‐derived HMGB1 released into the TME suppressed the host anti‐tumor immunity.

To clarify whether the inhibition of host anti‐tumor immunity observed for HMGB1 is a common phenomenon in tumor‐derived DAMPs, in our present study we focused on the effect of mitochondrial N‐formylpeptides, which have intracellular distributions and receptors completely different from those of HMGB1. We established MTFMT‐KO clones from murine tumor cells to deplete mitochondrial N‐formylpeptides in the tumor cells. The role of tumor cell‐derived mitochondrial N‐formylpeptides in anti‐tumor immunity was further investigated using the MTFMT‐KO tumor cells, and the following results were obtained:

(1) In vivo tumor growth of the MTFMT‐KO clones was markedly suppressed.

(2) The suppression was reversed by i.t. injection of mMT‐FPs.

(3) The suppression was mainly mediated by host T‐cell immunity.

(4) There was a marked increase in infiltration of macrophages and dendritic cells into the MTFMT‐KO tumor tissues.

(5) Infiltrations of CD4 cells, CD8 T cells and granulocytes into the MTFMT‐KO tumor tissues were also increased.

(6) Infiltration of PMN‐MDSCs into the MTFMT‐KO tumor tissues was decreased.

(7) The infiltration of MDSCs and M2 macrophages were increased by i.t. injection of mMT‐FPs in both WT and MTFMT‐KO tumors.

These results suggested that tumor‐derived mitochondrial formylpeptides promoted tumor growth through MDSC and M2 macrophage‐mediated inhibition of host T‐cell immunity. Therefore, inhibition of host anti‐tumor immunity by tumor‐derived DAMPs is a common phenomenon, at least for HMGB1 and mitochondrial formylpeptides.

Three different FPR have been identified. ⁶ , ²⁰ FPR1 is a high‐affinity receptor for the N‐formylpeptides and is expressed on the cell surface of neutrophils, macrophages, monocytes, and dendritic cells, as well as on the surface of non‐myeloid cells such as epithelial cells, hepatocytes, glial cells, astrocytes, and keratinocytes. ⁶ FPR2 is a low affinity receptor for N‐formylpeptides and also binds to ligands such as serum amyloid A, lipoxin A4, and Annexin A1; it is expressed on neutrophils, macrophages, monocytes, dendritic cells, microglia, T cells, and epithelial cells. ⁶ FPR3 was identified as a homologue of FPR1, but FPR3 is insensitive to N‐formylpeptide stimulation and its function is poorly understood. ⁶ In humans, expression of FPR3 is found in macrophages, monocytes, dendritic cells, and eosinophils. ⁶ In mice, the Fpr3 gene encodes the FPR2 protein and there is no FPR3 at the protein level. ²⁰

The clinical efficacy of cancer immunotherapy including ICI is largely influenced by the TME. ²¹ , ²² Our present experiments on mitochondrial N‐formylpeptides and previous results on HMGB1 ¹⁹ suggested that tumor‐derived DAMPs rendered the TME immunosuppressive. Therefore, if the immunosuppressive effects of the DAMPs could be blocked or removed from the TME by antibodies and/or antagonists, the clinical efficacy of the cancer immunotherapy may improve, and favorable outcomes could become achievable even in patients unresponsive to immunotherapy. The importance of subsequent induction of anti‐tumor immunity after treatment of cytotoxic regents or radiation therapy is well established. ²³ Therefore, the development of such antibodies or antagonists is important not only for immunotherapy but also for most cancer treatments. For mitochondrial N‐formylpeptides, development of such antibodies or antagonistic regents may also be useful for the treatment of aseptic SIRS induced by trauma or cardiac surgery.

In conclusion, we found that tumor cell‐derived mitochondrial N‐formylpeptides played a negative role in the host anti‐tumor immunity by modifying the tumor microenvironment. Further development of blockades of mitochondrial N‐formylpeptides, such as antibodies or antagonists, may be important not only for immunotherapy but also for most cancer treatments and the treatment of other inflammation‐related diseases.

AUTHOR CONTRIBUTIONS

Kayoko Waki: Formal analysis; investigation; validation; writing – original draft; writing – review and editing. Miyako Ozawa: Formal analysis; investigation; writing – review and editing. Keisuke Ohta: Investigation; methodology; writing – review and editing. Nobukazu Komatsu: Formal analysis; investigation; writing – review and editing. Akira Yamada: Conceptualization; formal analysis; funding acquisition; investigation; supervision; writing – original draft; writing – review and editing.

FUNDING INFORMATION

This study was supported by a JSPS KAKENHI Grant [no. JP23K06755 to AY].

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

ETHICS STATEMENTS

Approval of the research protocol by an Institutional Reviewer Board: N/A.

Informed Consent: N/A.

Registry and the Registration No. of the study/trial: N/A.

Animal Studies: All animal experimental protocols were approved by the Institutional Animal Care and Use Committee of Kurume University (approval no. 2019–030) in accordance with the national guidelines for the care and use of laboratory animals before starting the study. All gene modification experimental protocols were approved by the Institutional Genetic Modification Safety Committee of Kurume University (approval no. 31–16) in accordance with the national guidelines for research involving recombinant DNA experiments.

Supporting information

Figure S1.

Waki K, Ozawa M, Ohta K, Komatsu N, Yamada A. Tumor‐derived mitochondrial formyl peptides suppress tumor immunity through modification of the tumor microenvironment. Cancer Sci. 2024;115:3218‐3230. doi: 10.1111/cas.16266