Syntaphilin Regulates Neutrophil Migration in Cancer

Pathologically activated neutrophils (PMN) with immunosuppressive functions, termed myeloid-derived suppressor cells (PMN-MDSC), have emerged as critical contributors to tumor progression, including formation of premetastatic niches and regulation of immune responses in cancer (1–4). Migration of PMN-MDSC to distant sites is a critical component of their ability to promote tumor metastasis. Previously, we found that during early stages of tumor progression bone marrow (BM) PMN display potent ability to spontaneously migrate to distant sites and facilitate tumor metastasis (5). These cells lack the full biochemical and functional characteristics of MDSC but have upregulated genes associated with endoplasmic reticulum stress response and elevated rates of oxidative phosphorylation and glycolysis (5). However, the molecular mechanisms that regulate the increased migratory ability of the BM PMN have yet to be defined.

Leukocyte migration requires the generation of a polarized morphology characterized by the formation of leading-edge pseudopods where F-actin is concentrated, and a highly contractile cell rear known as the uropod where most cellular organelles and the microtubule organizing center are located (6, 7). This process requires energy in the form of ATP. Mitochondria are the most efficient source of ATP. Mitochondria cluster sites of high ATP demand in many cell types, and previous studies have suggested a possible direct functional interaction between these ATP-producing organelles and ATP-consuming cellular structures (8). Although PMN have relatively low abundance of mitochondria and historically the mitochondria in PMN were considered as largely to be involved in apoptosis (9), more recent studies have demonstrated a role for mitochondria in many PMN functions, including chemotaxis and antibacterial activity (10, 11). Furthermore, chemokine-induced redistribution of mitochondria toward the cellular uropod can activate actomyosin contraction during lymphocyte migration (12).

Rapid rearrangement of the mitochondrial network to facilitate the supply of energy for specialized functions has been described previously in several cell types (13). Syntaphilin (SNPH) is a cytoskeletal protein that tethers and immobilizes axonal mitochondria to microtubules (14). Overexpression of SNPH reduces the number of motile mitochondria in axons (15). We previously found that tumor progression is associated with downregulation or loss of SNPH in tumor cells, and this correlates with shortened patient survival, increased mitochondrial trafficking to the cortical cytoskeleton, greater membrane dynamics, and heightened cell invasion (16). Loss of SNPH in tumor cells or expression of an SNPH mutant lacking the mitochondrial localization sequence results in increased metastatic dissemination in xenograft or syngeneic tumor models in vivo. Tumor cells that acquire the ability to metastasize in vivo constitutively downregulate SNPH and exhibit higher oxidative stress, reduced cell proliferation, and increased cell motility (17). In this study, we tested the hypothesis that SNPH is involved in regulation of myeloid-cell migration. We identified a regulatory role for SNPH in PMN migration that directly affected tumor metastasis. SNPH deficiency increased PMN migration by increasing the speed and distance travelled by mitochondria through a mechanism involving phosphorylation of actomyosin, elevated metabolic activity of PMN, and increased production of adenosine. This mechanism is different to the mechanism by which SNPH deficiency enhanced mitochondria mobility in tumor cells and helps to understand increased motility of PMN in cancer with possible implications for therapeutic targeting.

Samples of peripheral blood were collected from 12 patients with previously untreated stage II–IV non–small cell lung cancer being cared for at the Helen F. Graham Cancer Center. Samples were collected before the start of treatment and processed on the day of collection. This cohort included 8 females and 4 males 48 to 74 years of age. Twelve healthy donors were enrolled. The study was conducted in accordance to ethical standards of the Declaration of Helsinki and was approved by the Institutional Review Board (IRB) of the Christiana Care Health System at the Helen F. Graham Cancer Center and The Wistar Institute. All patients and healthy donors signed IRB-approved consent forms.

Mouse experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of The Wistar Institute or AstraZeneca. BALB/c and C57BL/6 mice (female, 6–8 weeks old) were obtained from Charles River Laboratories, OT-I TCR transgenic mice [C57BL/6-Tg(TcraTcrb)1100Mjb/J, RRID:IMSR_JAX:003831; female, 6–8 weeks old], Pmel mice [B6.Cg-Thy1^a/Cy Tg(TcraTcrb)8Rest/J, PRID: IMSR_JAX:005023], and CD45.1 mice (B6.SJL-Ptprca Pepcb/BoyJ, RRID:IMSR_JAX:002014) were purchased from The Jackson Laboratory. Mice with a targeted deletion in the Snph gene Snph^tm1Vlcg on a C57BL/6N background were described previously and are referred to here as Snph-knockout (SNPH-KO) mice (18). Mice were obtained from Mutant Mouse Resources and Research Centers. Transgenic Ret melanoma (19) and KPC pancreatic carcinoma tumor models (20) were described previously. BM chimera were generated by intravenous transplantation of 5×10⁶ BM cells from wild-type (WT) or SNPH KO into lethally irradiated (1,000 rad) CD45.1 host mice. After 8 weeks, mice were bled to check reconstitution by flow cytometry for CD45.1 and CD45.2 and only mice with >87% chimerism were used for the experiments.

The LL2 (Lewis lung carcinoma) tumor cell line expressing luciferase was a gift from R. Ramakrishnan (H. Lee Moffitt Cancer Center) received in 2014. EL-4 (ATCC Cat# TIB-39, RRID:CVCL_0255), B16F10 (ATCC Cat# CRL-6475, RRID:CVCL_0159), and Lewis lung carcinoma (LLC; ATCC Cat# CRL-1642, RRID:CVCL_4358) cell lines were purchased from the ATCC in 2018. No additional authentication of the cell lines was performed. Cell lines were used for up to 8 passages and tested for Mycoplasma contamination by using the Universal Mycoplasma detection kit (ATCC, Cat#30–1012K). These cells were maintained in DMEM (Sigma-Aldrich, Cat#56436C) supplemented with 10% FBS (Sigma-Aldrich, Cat#12103C) and 1% penicillin–streptomycin (Thermo Fisher Scientific Cat#15140122) at 37°C with 5% CO₂.

A total of 0.5×10⁶ LL2-luciferase cells were injected subcutaneously. Three weeks later Luciferin (PerkinElmer, Cat#122799) was injected intraperitoneally in mice before the imaging according to the manufacturer's instruction. Mice were sacrificed, and bioluminescent images of the harvested lung were taken with IVIS Spectrum In Vivo Imaging System (IVIS, PerkinElmer) to detect luminescent signal coming from cancer cells. Analysis and quantification were performed with IVIS imaging system Living Image Software 4.7.4 (PerkinElmer). Photon counts in the area of interest were reported after subtraction of background.

A list of key reagents is provided in Supplementary Table S1.

For the isolation of BM Ly6G⁺ neutrophils, cells were labeled with anti-Ly-6G MicroBeads UltraPure (Miltenyi Biotec Cat#130–120–337) and separated on MACS columns (Miltenyi Biotec Cat#130–042–401). For isolation of splenic neutrophils, spleens were put in a 70-μm strainer placed on a conical 50 mL Falcon tube and cutoff into small pieces. These pieces are then grinded against the cell strainer using the plunger of a 5-mL syringe and washed several times with cold CSB. Tubes were then centrifugated at 1,500 rpm at 4°C, the supernatant was removed and red blood cells were lysed by resuspending the cell pellet in ammonium chloride lysis buffer for 5 minutes at room temperature. Single-cell suspensions from lungs were prepared using mouse lung dissociation kit (Miltenyi Biotec, Cat#130–095–927) according to the manufacturer's recommendations with an additional red blood cell lysis step as described above. Ly6G⁺ cells were then isolated using anti-Ly-6G MicroBeads UltraPure as described above.

CD11b⁺Ly6C^hiLy6G⁻ monocytes were sorted from BM of mice using MoFlo Astrios high-speed cell sorter.

The migration of unstimulated neutrophils or monocyte and chemotaxis of cells was measured as previously reported (5). Briefly, advanced RPMI (Thermo Fisher Scientific, Cat#12633–012) with the chemoattractants CXCL1 (BioLegend Cat#573702), CCL2 (R&D Systems, Cat#479-JE), or N-Formylmethionyl-leucyl-phenylalanine (fMLP; Sigma-Aldrich Cat#F3506) was placed in the bottom of the 3-μm-pore Transwell system (Neuro Probe Cat#101–3). Neutrophils were incubated with PSB12379 (Tocris) or 8-cyclopentyl-1,3-dipropylxanthine (DPCPX; Sigma-Aldrich, Cat#C101) in advanced RPMI for 10 minutes before being plated on top of the Transwell (0.1×10⁶ cells) and were incubated at 37°C, 5% CO₂ for 1 hour. 10-μL medium was taken from the bottom wells and cells were counted by a hemocytometer. The quantification of migrated neutrophils was done with the formula N = n×10⁴×0.029 (reflective of 29-μL medium in the bottom well).

Metabolic rates were determined using Seahorse XFe96 Flux Analyzers (Agilent Technologies) according to the manufacturer's protocol. In brief, the microplate was coated with 22.4 μg ml⁻¹ Cell-Tak (Fisher Cat#354240) using 200 mmol/L sodium bicarbonate. Approximately 400,000 neutrophils were seeded per well immediately after isolation 80 μL of unbuffered RPMI (Sigma-Aldrich; Cat#R6504) for XF96 analyzer. The microplate was incubated for 30 minutes at 37°C to allow the cells to settle into a monolayer. Unbuffered RPMI was gently added to the wells without disturbing the monolayer to bring the assay volume to 180 μL. The basal oxygen-consumption rate (OCR) and extracellular acidification rate (ECAR) were measured, in addition to rate changes after treatment with 5 μmol/L oligomycin (Sigma-Aldrich Cat#O4876), 1 μmol/L FCCP (Sigma-Aldrich Cat#C2920), and 0.75 μmol/L rotenone (Sigma-Aldrich Cat#R8875) and 1 μmol/L antimycin A (Sigma-Aldrich Cat#A8674).

Neutrophils from naïve tumor-free mice and PMN-MDSC from tumor-bearing mice were isolated from the BM, as described in Isolation of mouse neutrophils, and resuspended in culture medium that contains nutrients at physiological concentrations as was previously described (21). Cells were incubated for 1 hour in the presence of 5.5 mmol/L of ¹³C₆-glucose (Sigma-Aldrich, Cat#389374) and then harvested. For harvesting, cell pellets were washed twice in ice-cold PBS and extracted in a solution of LC/MS grade methanol, acetonitrile, and ultrapure water. Samples were centrifuged and the resulting cleared supernatant was transferred to a silanized MS vial and run by LC/MS.

LC/MS analysis was performed on a Q Exactive Hybrid Quadrupole-Orbitrap HF-X MS (Thermo Fisher Scientific) equipped with a HESI II probe and coupled to a Vanquish Horizon UHPLC system (Thermo Fisher Scientific). A total of 0.002 mL of sample was injected and separated by HILIC chromatography on a ZIC-pHILIC 2.1-mm column. Samples were separated by ammonium carbonate, 0.1% ammonium hydroxide, pH 9.2, and mobile phase B was acetonitrile. The LC was run at a flow rate of 0.2 mL/min and the gradient used was as follows: 0 minutes, 85% B; 2 minutes, 85% B; 17 minutes, 20% B; 17.1 minutes, 85% B; and 26 minutes, 85% B. The column was maintained at 45°C and the mobile phase was also preheated at 45°C before flowing into the column. The relevant MS parameters were as listed: Sheath gas, 40; auxiliary gas, 10; sweep gas, 1; auxiliary gas heater temperature, 350°C; spray voltage, 3.5 kV for the positive mode and 3.2 kV for the negative mode. Capillary temperature was set at 325°C, and funnel RF level at 40. Samples were analyzed in full MS scan with polarity switching at scan range 65–975 m/z; 120,000 resolution; automated gain control target of 1E6; and maximum injection time of 100 ms. Identification and quantitation of metabolites was performed using an annotated compound library and TraceFinder 4.1 software for LC/MS data processing. Metabolite measurements were normalized based upon protein concentration determined from cell pellets.

Total BM cells were stained with AQUA, CD11b-BV421 (Supplementary Table S1), and Ly6G-APC (BD Bioscience Cat#560599). Cells were washed and stimulated with doses of CXCL1 and fMLP in Advanced RPMI for the time points indicated in the figure legends. Cells were immediately fixed and permeabilized using Cytofix/Cytoperm Solution (BD Biosciences Cat#554722) and washed in Perm/Wash Buffer (BD Biosciences) following the manufacturer's protocol. Cells were stained with 1 U of Phalloidin-AF488 (Thermo Fisher Scientific Cat#A12379) for 20 minutes at 4°C, washed in Perm/Wash Buffer, run on a LSRII flow cytometer, and analyzed using FlowJo (Tristar version 10.8.0).

Neutrophils were isolated as described in Isolation of mouse neutrophils using Ly6G beads with purity was >95%. Total RNA was extracted from cell pellets using the Direct-zol RNA Miniprep (Zymo Research, Cat#R2051). RNA quality was validated using the TapeStation RNA ScreenTape (Agilent, Cat#5067–5579). 100 ng of total RNA was used to prepare a library for Illumina Sequencing using the Quant-Seq 3′mRNA-Seq Library Preparation Kit (Lexogen, Cat#015.96). Library quantity was determined using qPCR kit and absolute quantification with standard curve method (KAPA Biosystem). Overall library size was determined using the Agilent TapeStation and the DNA High-Sensitivity D5000 ScreenTape (Agilent, Cat# G2992AA). Equimolar amounts of each sample library were pooled, denatured and Mid-Output, Paired-End, 150 cycle Next-Generation Sequencing was done on a NextSeq 500 (Illumina). Three samples of WT and SNPH-deficient BM neutrophils were evaluated. No technical replicates were used. Lexogen Quantseq Library Preparation Kit was used for library preparation. Single-end read runs were used, with read lengths up to 75 bp in high output mode, with 14M per sample total read counts. Raw data were aligned against mm10 genome and gene-level read counts and FPKM values were estimated with Ensemble transcriptome information using RSEM v1.2.12 software (22). SNPH KO versus WT sample groups were compared using DESeq2 (23) software and normalized by the DESeq2 values were used for heat map visualization. Overall changes were considered significant if passed FDR<5% threshold. Gene set enrichment analysis for genes that passed nominal P < 0.05 threshold was performed using QIAGEN's Ingenuity Pathway Analysis software (IPA, QIAGEN Redwood City, www.qiagen.com/ingenuity) based on “Canonical Pathways” and “Networks” options. RNA sequencing (RNA-seq) data were deposited to NCBI GEO data repository under accession number GSE190505.

RNA was extracted from mouse and human neutrophils using the Quick-RNA Miniprep Kit (Zymo Research Cat#R1055) according to the manufacturer's instructions using 100 ng of the template. Reverse transcription was generated with the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems Cat#4374967). Quantitative PCR was then performed using Power SYBR Green PCR Master Mix (Applied Biosystems Cat#4368708) on an ABI 7500 Fast instrument. Each experiment was performed in triplicates. Expression of mouse genes; Snph, Mfn1, Mfn2, and Drp1 was normalized to the expression of 18s and measured using ΔΔC_t. Expression of human SNPH gene was normalized to GAPDH and measured using ΔΔC_t. Primers are described in Supplementary Table S2.

PMN-MDSC and neutrophils were isolated from peripheral blood using double-density gradient of Histopaque-1077 and Histopaque-1119 (Sigma-Aldrich Cat#10771 and Cat#11191, respectively). Then CD15-beads (Miltenyi, Cat#130–046–601) were used to isolate PMN-MDSC from the low-density peripheral blood mononuclear cells and neutrophils from the high-density gradient. The purity of neutrophils and PMN-MDSC with CD14⁻CD15⁺ phenotype was checked by flow cytometry and was >95%.

Antibody incubations were performed at 4°C for 15 minutes. Usually, up to 1×10⁶ cells were incubated with Fc-block (BD Biosciences, Cat#553142) for 10 minutes and surface staining was performed at 4°C for 15 minutes. For intracellular staining, cells were then fixed with FOXP3/transcription factor staining buffer set (eBioscience Cat#00–5523–00) according to the manufacturer's instructions and incubated for 45 minutes at 4°C. Samples were analyzed using an LSRII or Symphony flow cytometer (BD Biosciences) and data were analyzed by FlowJo software 10.8.0 (Tristar RRID:SCR_008520). A list of antibodies used is in Supplementary Table S1. Gating strategy is shown in Supplementary Fig. S1.

Assessment of mitochondrial membrane potential was performed by staining with tetramethylrhodamine ethyl ester (TMRE; Thermo Fisher Scientific, Cat#T669; 20 nmol/L) in complete seahorse media (2 mmol/L glutamine and 20 mmol/L glucose) for 20 minutes at 37°C. Excess mitochondrial dye was removed by washing and cells were further acquired by flow cytometry.

To measure proliferation, CD8⁺ T cells were purified from spleens from WT and SNPH KO mice using easysep Mouse CD8⁺ T-cell isolation Kit (Stemcell Cat#19853), stained with Cell trace Far Red (Thermo Fisher Scientific, Cat#C34564) and cultured with anti-CD3/CD28 beads (Thermo Fisher Scientific, Cat#11452D) and IL2 (100 U/mL; PeproTech Cat#212–12) for 3 days. After 3 days, T-cell proliferation was assessed by flow cytometry. To measure IFNγ, CD8⁺ T cells were activated with anti-CD3/CD28 beads and IL2 (100 U/mL) then restimulated with PMA/IONO [eBioscience Cell Stimulation Cocktail (plus protein transport inhibitors; 500X) Cat#00–4975–93] for 6 hours and intracellular IFNγ was measured by flow cytometry.

Neutrophils from BM were lysed in RIPA buffer (Sigma-Aldrich Cat#R0278) in the presence of protease inhibitor cocktail (Sigma-Aldrich Cat#P8340) and stored at −80°C. Whole-cell lysates were prepared and subjected to 10% SDS–PAGE and transferred to polyvinylidene difluoride membrane. The membranes were probed overnight at 4°C with primary antibody. Membranes were then washed and incubated for 1 hour at room temperature with secondary antibody conjugated with peroxidase and visualized with ECL (Amersham Biosciences, Cat# RPN2232). A list of antibodies used is in Supplementary Table S1.

Changes in mitochondrial morphology were investigated as described previously (24). High-optical-quality glass-bottom 35-mm plates (MatTek Corporation) were precoated with poly-L-lysine (Sigma-Aldrich, Cat#P4707) for 30 minutes at 37°C. WT and SNPH-KO PMN (1 × 10⁴) were seeded and stained with 100 nmol/L MitoTracker Deep Red (Thermo Fisher Scientific, Cat#M46753) for 30 minutes and washed with complete medium. Cells were imaged on a Leica TCS SP8 X inverted confocal laser scanning microscope using a 63×, 1.40-numerical-aperture oil objective. Live imaging time-lapse was performed using a Tokai Hit incubation chamber equilibrated to bidirectional scanning at 8,000 Hz at 37°C with 5% CO₂. Images were acquired every 3 seconds for 1 minute. Individual 12-bit images were acquired using a white-light supercontinuum laser (0.2% at 633 nm) and hybrid detectors at a 5× digital zoom with a pixel size of 70 nm by 70 nm and a step size of 0.260 μmol/L. At least 6 single cells under each condition were collected for analysis. Initial postprocessing of the 3D sequences was carried out with Hyugens software to deconvolve the images. Images were imported into Leica LAS X software to study fission and fusion events and mitochondrial movements.

The wells of μ-Slide 4 Well Ph+ (ibidi) were coated with 50 μg/mL fibronectin (Sigma-Aldrich, Cat#F1141). Cells were plated (0.5 × 10⁶ cells per well) in advanced RPMI without FBS and were incubated at 37°C and 5% CO₂ for 10 to 15 minutes to allow cell attachment. The wells were gently washed with 2 × PBS to remove nonadherent cells, and Advanced RPMI was added to the wells. Cells were placed onto a motorized stage and were observed with a Nikon Eclipse TE300 fluorescence microscope maintained in an environment of 37°C and 5% CO₂. A 10 × or 20 × objective was used to capture images during the course of the time lapse. Images were captured every 30 seconds over the course of 30 minutes from at least 6 different fields of view. Cell movement was tracked and analyzed using the ImageJ plugin Manual Tracking and the Chemotaxis tool, respectively. ImageJ and the plugin are both freely available through the NIH website (http://rsbweb.nih.gov/ij/).

BM Ly6G⁺ neutrophils were prepared as described in Isolation of mouse neutrophils. Then cells were plated in U-bottomed 96-well plates in triplicates in complete RPMI and cocultured at various ratios with total splenocytes from Pmel or OT-I mice in the presence of cognate peptides (OT-I, SIINFEKL, Anaspec Cat#AS-60193–1; Pmel, EGSRNQDWL, Cat# Anaspec AS-64752) for 48h. Then [³H] thymidine (PerkinElmer, Cat#NET027×001MC) was added (1 μL/well), followed by incubation overnight. Proliferation was measured with a TopCount NXT instrument (PerkinElmer).

Statistical analysis was done using unpaired two-tailed Student t tests, two-way ANOVA with correction for multiple comparisons, or Fisher Exact Test. Statistical tests were performed using GraphPad Prism version 8.0 (GraphPad Prism; RRID:SCR_002798). P values of 0.05 were considered significant.

RNA-seq data were deposited to NCBI GEO data repository under accession number GSE190505. All other data are available in the main text or the Supplementary Materials or are available upon request from the corresponding author.

Naïve, tumor-free SNPH-KO mice had a substantially lower proportion of CD11b⁺Ly6C^loLy6G⁺ PMN among CD45⁺ cells in BM and spleens than WT mice. No differences in the proportion of CD11b⁺Ly6C^hiLy6G⁻ monocytes (MON) were observed (Fig. 1A). Although no differences in the proportion of PMN and MON in peripheral blood between WT and SNPH-KO mice were found (Supplementary Fig. S2A), we observed markedly higher proportion of PMN among CD45⁺ cells in lungs of SNPH-KO mice than in WT mice (Fig. 1B). The proportions of MON and dendritic cells were not different, whereas the proportion of macrophages was decreased (Fig. 1B). We investigated whether differences in the proportions of PMN in BM and lungs of WT and SNPH-KO mice could be the result of differences in the survival of these cells. To address this question, we isolated PMN from BM, spleen, and lung of WT and SNPH-KO mice and cultured them overnight in complete medium without cytokines. The absence of SNPH did not affect the proportion of recovered cells Supplementary Fig. S2B.

Taken together, these results suggested a possible effect of SNPH on the migration of PMN to the tissues. To address this possibility directly, we evaluated migration of PMN and MON in a standard Transwell membrane assay. BM PMN but not BM MON from SNPH-KO mice had markedly higher spontaneous migration than their WT counterparts (Fig. 1C). Increased spontaneous migration of SNPH-deficient PMN was confirmed by time-lapse microscopy (Fig. 1D and E).

We investigated whether SNPH deletion would affect PMN presence in the lungs of tumor-bearing mice. To address this question, mice were injected subcutaneously with 0.5×10⁶ LL2 cells, which are known to spontaneously metastasize to lungs. The presence of PMN and MON was evaluated three weeks after tumor injection (tumor size >1.5 cm in diameter). We observed large accumulation of PMN in lungs of WT and SNPH-KO tumor-bearing mice (Fig. 1F), with no statistical difference. These results support the concept that once tumor is established and produces a large number of chemokines (as could be the case in 3-week LLC tumors with propensity to metastasize to the lung), the effect of SNPH deletion is lost.

To further investigate SNPH involvement in PMN migration, we assessed PMN migration in response to chemotactic signals. Although PMN from SNPH-KO mice migrated better than PMN from WT mice in response to fMLP, no increase in the ratio of the responses to the chemokine was found (Fig. 2A). PMN from SNPH-KO mice migrated better than PMN from WT mice in response to the chemokine CXLC1. However, as was the case with fMLP, no changes in the ratio of response were observed (Fig. 2B). Expression of CXCR2, which is the main receptor responsible for CXCL1-mediated chemotaxis, was the same in PMN from WT and SNHP-KO mice (Fig. 2C). Although no statistically significant differences in spontaneous migration of MON between WT and SNPH-deficient cells was observed, we tested the possibility that the response of MON to chemokines could be different. However, MON from WT and SNPH-KO mice responded equally well to stimulation with the monocyte-specific chemokine CCL2 (Fig. 2D). Thus, deletion of SNPH enhanced spontaneous migration of PMN, which also resulted in higher migration in the presence of chemokines.

Next, we studied the biological relevance of SNPH expression in PMN. We evaluated Snph expression in BM CD11b⁺Ly6C^loLy6G⁺ PMN from transgenic and transplantable tumor models that previously demonstrated different pattern of PMN migration (5). BM PMN collected from genetically engineered RET melanoma and pancreatic cancer KPC models (after tumors were developed) had substantially reduced expression of Snph. In contrast, BM PMN from subcutaneously injected EL4 and LLC tumor-bearing mice showed no difference in Snph expression from naive tumor-free mice (Fig. 3A). These data paralleled the differences in BM PMN migration previously observed in mice bearing transgenic (high spontaneous PMN migration) and subcutaneous transplantable models (no differences in migration activity from control PMN; ref. 5). PMN from patients with cancer had substantially lower expression of SNPH (Fig. 3B) than PMN from healthy donors. In paired samples from the same patients, PMN had lower expression of SNPH than PMN-MDSC (Fig. 3C). These results were also consistent with the pattern of migration of PMN and PMN-MDSC from patients with cancer described previously showing substantially higher migration of PMN from patients with cancer than PMN-MDSC (5).

Next, we investigated the effect of SNPH deletion on tumor metastases. Luciferase-expressing LL2 was established subcutaneously in WT and SNPH-KO mice. Mice were evaluated three weeks after tumor inoculation when tumor size reached 1.5 cm in diameter. No differences in size of primary tumors were observed (Fig. 4A). In contrast, SNPH-KO mice had significantly more lung metastases than WT mice (Fig. 4B). To further investigate a possible effect of SNPH deletion in hematopoietic cells on metastatic spread, we established BM chimeras by reconstituting lethally irradiated WT congenic mice with BM cells from WT or SNPH-KO mice. Six weeks after reconstitution and confirmation of >87% of donor hematopoietic cell phenotype (Fig. 4C), we evaluated the growth of LL2 tumors established subcutaneously. No differences in growth of primary tumors between mice reconstituted with WT and SNPH-KO BM were observed (Fig. 4D). However, mice reconstituted with BM from SNPH-KO donors showed marked increase in the presence of lung metastases. Whereas only 26% of mice reconstituted with WT BM had developed lung metastasis 3 weeks after subcutaneously tumor inoculation, tumor lesions in lungs were found in 60% of mice reconstituted with SNPH-KO BM (P = 0.022; Fig. 4E).

To validate these observations, we tested a different tumor model. WT and SNPH KO mice were challenged subcutaneously with B16F10 melanoma. Three weeks later lung metastasis was evaluated. We observed a significantly higher number of lung metastases in SNPH KO mice than in WT mice (Fig. 4F). Because these results suggested that the deletion of SNPH promoted spontaneous migration of PMN to tissues and tumor metastasis, we investigated whether this would also affect the initial development of the tumors. To test this hypothesis, we used BM chimeras, established as described above. However, in these experiments we injected small amount of tumor cells—7×10⁴ cells. In mice reconstituted with WT BM no tumors were visible four weeks after injection in all 19 mice tested. In contrast, 15 out of 20 mice reconstituted with SNPH-KO BM had developed tumors (Fig. 4G).

Next, we tested the possibility that SNPH could affect the suppressive activity of PMN-MDSC. Splenic PMN-MDSC from WT and SNPH-KO LL2 tumor-bearing mice demonstrated similar levels of immune suppressive activity (Fig. 4H). SNPH deletion did not alter the proportion of PMN in BM or spleens (Supplementary Fig. S2C). There was a trend in decreased presence of MON but it did not reach statistical significance (Supplementary Fig. S2D). In the Transwell assays, spontaneous migration of PMN-MDSC from SNPH-KO mice bearing LL2 tumors was markedly higher than from WT mice (Fig. 4I).

We also assessed the effect of SNPH deletion on T cells. No difference in the presence of CD8⁺ T cells, CD4⁺FOXP3⁺ Treg cells, and CD4⁺Foxp3⁻ T cells was found between WT and SNPH-KO mice (Supplementary Fig. S3A–S3C). Stimulation of CD8⁺ T cells with anti-CD3/anti-CD28 resulted in equivalent proliferation of WT and SNPH-deficient cells (Supplementary Fig. S3D). Similarly, no differences were observed in IFNγ production by T cells in response to PMA/Ionomycin stimulation (Supplementary Fig. S3E). Thus, the data suggest that increased migration of SNPH-deficient PMN can contribute to tumor metastases.

To understand the possible mechanisms by which SNPH regulates PMN, we performed whole-transcriptome RNA-seq of BM PMN isolated from WT and SNPH-KO mice. 113 genes were found to be significantly differentially expressed (FDR < 5%). Among those genes, 48 were upregulated in PMN from SNHP-KO mice and 65 were downregulated. The 25 most changed genes are shown in Supplementary Fig. S4A. Cellular assembly and organization, cellular function and maintenance, energy production, and free radical scavenging were among the most changed gene networks (Supplementary Fig. S4B). Paxillin and ephrin receptor signaling, which are involved in cell motility, were among the most activated pathways in SNPH-deficient PMN (Supplementary Fig. S4C). These differences suggested that signaling involved in cell motility is the biggest effect of loss of SNPH.

Neutrophil migration requires Rac-1–dependent actin polymerization at the leading edge to push the front of the cell forward and Rho-dependent actomyosin contraction at the uropod to retract the trailing edge (6, 25). We evaluated actin polarization (F-actin) in SNPH-deficient PMN. Stimulation with CXCL1 resulted in an increase in F-actin. However, we found no difference between PMN from WT and SNPH-KO mice (Fig. 5A). The retraction of the cell membrane at the uropod is accomplished through contractile forces generated by myosin motor proteins and the consequent interactions with actin filaments at lateral and posterior regions of the cell cortex (6). This process is regulated by Rho-A–mediated phosphorylation of myosin light chain2 (MLC2). We found that PMN from SNPH-KO mice had substantial upregulation of pMLC2 relative to WT PMN (Fig. 5B).

How can deficit of SNPH enhance phosphorylation of MLC2 and PMN migration? Allocating mitochondria to the uropod of a migrating cell can provide ATP to sustain actomyosin contraction (12). We observed that BM PMN from SNPH-KO mice had a markedly higher OCR than WT PMN, which is indicative of increased oxidative phosphorylation (Fig. 5C). Those changes were associated with a trend toward an increase in ATP production (Fig. 5D). Consistent with a greater capacity for oxidative phosphorylation, PMN from SNPH-KO mice had increased expression of mitochondrial complex I, but not complex II subunits succinate dehydrogenase A (SDHA) and B (SDHB) compared with PMN from WT mice (Fig. 5E). Furthermore, the mitochondrial membrane potential, evaluated by staining with tetramethylrhodamine methyl ester (TMRM), was increased in SNPH-deficient PMN (Fig. 5F). Consistent with these findings, SNPH deletion did not affect cytoplasmic- or mitochondria-derived superoxide production by PMN nor mitochondrial mass in these cells (Supplementary Fig. S5A–S5C).

A previous study demonstrated that downregulation of SNPH in tumor cells promoted cell motility by increasing the speed and distance travelled by individual mitochondria and facilitating organelle fusion and fission (16). Although slight upregulation of phosphorylated dynamin-related protein 1 in SNPH-deficient PMN (Fig. 5B) could suggest possible changes in mitochondria fission, direct evaluation of mitochondria fusion, and fission did not demonstrate significant differences (Fig. 5,G and H). This was consistent with a lack of difference in the expression of molecules involved in mitochondrial dynamics (Supplementary Fig. S5D). In contrast, mitochondria in SNPH-deficient PMN exhibited faster speed of movement and greater distance travelled compared with WT PMN (Fig. 5I; Supplementary Video S1). These data suggest that increased mitochondrial movement can facilitate contracting of actomyosin and facilitate PMN movement.

In addition to oxidative phosphorylation, PMN from SNPH-KO mice also had a higher ECAR indicative of increased glycolysis (Fig. 6A). To better assess changes in glycolysis in SNPH-deficient PMN, we performed metabolomics analysis of ¹³C₆-glucose through glycolysis. PMN from SNPH-KO mice had higher rates of incorporation of the ¹³C into the glycolytic intermediates glucose-6-phosphate, glyceraldhedyde-3-phosphate, 3-phosphoglycerate and phosphoenolpyruvate than WT PMN (Fig. 6B). Only slight increases in intracellular pyruvate and lactate were found in SNPH-deficient PMN (Fig. 6B). The ¹³C-labeling patterns strongly indicated that loss of SNPH caused greater flux of glucose carbons through glycolysis in PMN. More detailed analysis of metabolites is shown in Supplementary Fig. S6. Consistent with the result obtained during the analysis of oxidative phosphorylation, there was higher ATP synthesis by SNPH-deficient PMN as compared with WT PMN (Supplementary Fig. S6). In addition, we found a considerably higher amount of adenosine in SNPH-deficient PMN than in WT PMN (Fig. 6C). It has been long appreciated that adenosine, by activating the A₁ and A₃ receptors, promotes migration of PMN (26). We hypothesized that increased production of adenosine contributed to enhanced migration of SNPH-deficient PMN. The pathway of adenosine generation involves the hydrolysis of ATP by ectonucleotidases such as CD39 (ectonucleoside triphosphate diphosphohydrolase 1) and CD73 (5-nucleotidase; refs. 27, 28). We assessed the effect of inhibition of the generation of adenosine on the ability of SNPH-deficient PMN to migrate. WT and SNPH-deficient PMN were treated with PSB 12379, an inhibitor of CD73, and we observed that the spontaneous migration of SNPH-deficient PMN was completely abolished (Fig. 6D). To further validate these observations, we performed experiments with the selective adenosine A₁ receptor (A1AR) antagonist DPCPX. Inhibition of A1AR completely abrogated increased spontaneous migration of SNPH-deficient PMN (Fig. 6E). Thus, SNPH-deficient PMN had increased oxidative phosphorylation, TCA-cycle flux, and glycolysis, which resulted in modest increases in ATP and substantially greater production of adenosine than in PMN from WT mice, which resulted in increased migratory ability of these cells.

In this study, we described the association of SNPH loss with the enhanced ability of PMN to spontaneously migrate. Loss of SNPH increased the speed and distance travelled by individual mitochondria within neutrophils, elevated metabolic activity in neutrophils, which together promoting these cells migration.

Mitochondria have recently emerged as a key regulator of leukocyte migration. Chemoattractants induce the redistribution of mitochondria toward the uropod of polarized migrating leukocytes through a mechanism involving microtubules and mitochondrial fission (29). A study in tumor cells demonstrated that SNPH can downregulate mitochondrial fission and fusion, buffer oxidative stress and maintain complex II-dependent bioenergetics (30). Tumor cells that constitutively downregulate SNPH exhibit higher oxidative stress and increased cell motility (17). However, in PMN, no effect of SNPH deletion on mitochondrial dynamics was found. Herein, we have shown that elevated oxidative phosphorylation and glycolytic rates drove the increased migration of SNPH-deficient PMN. SNPH deficiency in PMN resulted in increased glycolysis and oxidated phosphorylation and there was no increase in mitochondria-produced superoxide. Although the precise mechanism of these differences is not clear, the explanation could be in the nature of mitochondrial involvement in PMN function. Under steady-state conditions, PMN do not rely on OXPHOS and mitochondria for their function and have a relatively small number of mitochondria. The main source of superoxide in PMN is NADPH oxidase. It is conceivable that increased mitochondrial movement as a result of loss of SNPH is sufficient to cause strong functional effect on PMN. In leukocytes, oxidative phosphorylation is required for mitochondrial-dependent migration (31). These findings suggest that increased metabolic activity could provide ATP to sustain actomyosin contraction, thereby enabling retraction of the trailing edge and cell advance.

Mitochondria supply ATP essential for synaptic transmission. In neurons, SNPH immobilizes mitochondrial transport and causes local energy deficits. Enhancing mitochondrial transport via genetic manipulation of SNPH facilitated regenerative capacity of axons by replenishing healthy mitochondria in injured axons, thereby rescuing energy deficits (32). Functional micro-compartmentalization of intracellular metabolites (including adenine nucleotide) and substrates have long been recognized. Our data suggest that accumulation of mitochondria at the uropod of a migrating cell might provide ATP in this strategic position, where class II myosin proteins are localized. Our data suggest a contribution of adenosine to increased migration of SNPH-deficient PMN. Intracellular adenosine is generated by stepwise dephosphorylation of ATP or by the hydrolysis of S-adenosylhomocysteine (33). Adenosine formed extracellularly can also be transported into cells via nucleoside transporters (34). Neutrophils can both release adenosine and respond to it via expression of all four adenosine receptor subtypes. At low concentrations, adenosine can act via the A₁ and A₃ adenosine receptor subtypes to promote neutrophil chemotaxis and phagocytosis (26). Thus, decrease of SNPH in neutrophils may facilitate their spontaneous migration via combined effect of increased local concentration of ATP and adenosine signaling.

It is widely recognized that an increased flux of neutrophils contributes to the formation of the premetastatic niche. Depletion of PMN and PMN-MDSC in mice decreases the formation of metastasis (2, 35, 36). Our data with SNPH-KO mice are consistent with these observations. PMN-MDSCs are not the only possible contributors to metastasis. Monocytes/macrophages are plying a major role in the development of tumor metastasis (37, 38). However, our data suggest that deletion of SNPH is probably unlikely to exert its effect via recruitment of monocytes and accumulation of macrophages. First, the number of monocytes in BM and spleen was only marginally affected by SNPH deletion, whereas the number of PMN was substantially reduced. Second, in lungs, in contrast with PMN, the presence of monocytes and macrophages was not different between SNPH-KO and WT mice. Third, migration of SNPH-deficient monocytes was only marginally affected.

We believe the reason why neutrophils accumulate in the tissues in SNPH-KO mice is that in the absence of SNPH, neutrophils acquire a potent ability to spontaneously migrate. Normal tissues produce baseline amount of chemokines, which attract neutrophils patrolling the tissues. We hypothesize that SNPH-deficient neutrophils become much more active in response to the low chemokine gradient in tissues, which causes the accumulation of PMN in tissues not yet affected by growing tumors. This is clinically relevant in early-stage tumors prone to metastasis. Downregulation of SNPH in PMN would promote their migration to unaffected tissues and the development of distant metastases.

Our data suggest one possible molecular mechanism that regulates this flux. Increased PMN migration to lung was associated with increased metastatic spread to lung. SNPH deficiency did not affect suppressive activity of PMN-MDSC, presence or function of T cells. Thus, SNPH loss may supply distant tissues with immune suppressive PMN-MDSC able to facilitate tumor metastases.

E. Sanseviero reports employment at AstraZeneca. E.L. Stone reports personal fees from GigaGen outside the submitted work. Y. Nefedova reports grants from NIH/NCI during the conduct of the study; grants and nonfinancial support from Active Biotech, Buzzard Pharmaceuticals, Merck, and Jubilant Therapeutics Inc. outside the submitted work. D.I. Gabrilovich reports other support from AstraZeneca outside the submitted work. No disclosures were reported by the other authors.

S. Fu: Formal analysis, investigation, methodology, writing–original draft, writing–review and editing. H. Deng: Investigation. I. Bertolini: Investigation, methodology. M. Perego: Formal analysis, investigation. E.S. Chen: Investigation, writing–review and editing. E. Sanseviero: Investigation. A. Mostafa: Investigation. K. Alicea-Torres: Investigation. L. Garcia-Gerique: Investigation. E.L. Stone: Formal analysis, supervision. A.V. Kossenkov: Formal analysis. Z.T. Schug: Formal analysis, investigation. B. Nam: Resources. C. Mulligan: Resources. D.C. Altieri: Funding acquisition, writing–review and editing. Y. Nefedova: Formal analysis, funding acquisition, project administration, writing–review and editing. D.I. Gabrilovich: Conceptualization, formal analysis, supervision, funding acquisition, writing–original draft, writing–review and editing.

This work was supported by NCI grants P01 (CA140043) and R35 (CA220446) as well as Wistar Cancer Center support grant CA010815 (to D.C. Altieri).

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