The carbohydrate sialyl Lewis X (sLeX) is expressed on leukocytes and carcinoma cells and binds to selectins during inflammatory processes and early metastasis. Synthesis of sLeX depends on activity of enzymes, including α(1,3/1,4) fucosyltransferase (FucT-III). Tumor necrosis factor-α (TNF-α) up-regulates FucT-III, resulting in increased sLeX in the airways of patients with respiratory disease; however, the mechanisms that regulate sLeX in the inflammatory tumor microenvironment are not well understood.
The authors stably transfected human lung carcinoma cell lines with the FucT-III gene and exposed them to TNF-α to investigate its role in regulation of sLeX expression and selectin-binding ability using semiquantitative real-time polymerase chain reaction and flow cytometry. Cytokine expression was examined in transfected cells using chemiluminescent arrays and enzyme-linked immunosorbent assays, and invasion was studied using Matrigel assays and alterations in morphology. Human lung tissue arrays were analyzed for immunohistochemical detection of sLeX and neutrophils.
Stimulation of FucT-III–transfected cells with recombinant human (rh) TNF-α up-regulated sLeX expression and increased E-selectin binding. Transfected cells secreted high levels of interleukin 8, growth-regulated oncogene-α, and mast cell proteinase-1. Cells exposed to rhTNF-α, neutrophil-conditioned media, and cultures with a 5:1 ratio of neutrophils to cancer cells had significantly increased sLeX expression and invasiveness and underwent nonadherent morphologic changes. In lung carcinomas, but not in normal lung tissues, 71% of tumors were highly positive for sLeX expression in areas of increased neutrophil infiltration.
Lung cancer is the leading cause of cancer-related deaths in the United States and worldwide. Of these cancers, 80% are nonsmall cell lung cancer (NSCLC), and mortality commonly is caused by locally invasive or recurrent metastatic disease.1 Alterations in tumor-associated carbohydrates occurs in many cancers and are highly correlated with decreased survival because of metastasis.2 Specifically, the carbohydrate sialyl Lewis X (sLeX) often is associated with advanced malignancies and a poor prognosis in many carcinomas.3 Increased postoperative distant metastasis and mortality are linked to higher expression of sLeX on NSCLC cells.4 SLeX is a ligand for the selectin adhesion molecules (E-selectin, P-selectin, and L-selectin), of which endothelial E-selectin is the major target. Selectins modulate the binding of circulating inflammatory cells and tumor cells to blood vessels during inflammatory processes and metastasis.5
The contribution of inflammation to metastasis in NSCLC is not well understood. Tumor-associated macrophages are believed to be the predominant myeloid cells that promote tumor progression,6 but controversial results have been obtained for the role of macrophages in NSCLC. High numbers of tumor-infiltrating macrophages reportedly were associated with shorter relapse-free survival of patients.7 Alternatively, others did not observe an association between macrophage count and outcome in patients with NSCLC.8 Our focus in the current study was to investigate the influence of tumor-associated neutrophils on sLeX expression and metastatic properties of NSCLC cells, because these factors may predict tumor advancement. Recent evidence suggests that neutrophils are protumorigenic, but their role in tumor progression is not well defined.9 Infiltration of neutrophils in lung cancers is associated with a poor outcome.10 It has been reported that tumor-elicited neutrophils increased tumor metastatic potential and invasiveness in a rat mammary adenocarcinoma model11 and may regulate invasion of tumors by activation of matrix metalloproteinases that remodel the extracellular matrix.12, 13 Tumor-associated macrophages and neutrophils may promote a metastatic phenotype through secretion of the proinflammatory cytokine tumor necrosis factor-α (TNF-α).14, 15 In contrast to its better-known anticancer effects, TNF-α also has protumorigenic activity in many cancers.16 In lung cancer, the role of TNF-α is not well understood: TNF-α reportedly up-regulated sLeX expression in a lung cancer cell line, and such up-regulation may be associated with increased metastatic potential.17
Our current study focused on NSCLC disease progression from a novel perspective. We investigated the interactions of neutrophils with NSCLC cells involving the tumor cell-surface carbohydrate sLeX and the contribution to malignant cellular invasion. Our results indicated that neutrophils induce sLeX expression on NSCLC cells and promote their nonadherent growth and invasiveness by a TNF-α–dependent mechanism. Our findings are applicable to the development of therapies to reduce NSCLC invasion and, ultimately, metastasis.
MATERIALS AND METHODS
Cell Culture and Stable Transfections
H1299 and A549 NSCLC cell lines were obtained from American Type Culture Collection (Manassas, Va) and were authenticated and cultured in recommended medium under standard conditions. The human α(1,3/1,4) fucosyltransferase (FucT-III) gene was removed from a pRc/RSV vector (a gift from Dr. Rodger P. McEver; Oklahoma Medical Research Foundation, Oklahoma City, Okla) and cloned into a pcDNA3.1(+)neo expression vector (Invitrogen, Carlsbad, Calif) using HindIII and XbaI restriction enzymes to create the FucT-III/pcDNA3.1(+)neo plasmid. The FucT-III/pcDNA3.1(+)neo and pcDNA 3.1(+)neo plasmids were linearized by digestion with the Mfe I restriction enzyme. H1299 and A549 lung carcinoma cells were cultured in a 6-well plate (3.5 × 105 cells per well) for 24 hours. Then, the cells were transfected with 4 μg of each linearized plasmid using Lipofectamine 2000 (Invitrogen) and incubated for 48 hours. Heterogeneous populations of stably FucT-III–transfected and mock-transfected cells were selected with 1 mg/mL G418 (Sigma-Aldrich, St. Louis, Mo) for 14 days before they were used in the experiments. Multiple aliquots of selected, stable transfectants also were stored in liquid nitrogen. Transfected cells were maintained in medium containing G418 for up to 1 month; then, the cells were discarded, and new aliquots were thawed for use. FucT-III gene activity was verified by analysis of messenger RNA (mRNA) levels by semiquantitative reverse transcriptase-polymerase chain reaction (RT-PCR) and analysis of sLeX expression by flow cytometry, as described below.
Flow Cytometric Analysis
The sLeX epitope on cells was detected using the monoclonal antibodies (MoAbs) (10 μg/mL) mouse antihuman immunoglobulin M (IgM) CSLEX-1 (generously provided by Dr. Bruce Walcheck, University of Minnesota) or an IgM isotype control (Caltag Laboratories, Burlingame, Calif). The secondary antibody was phycoerythrin (PE)-conjugated F(ab′)2 goat antimouse IgM (Jackson Immunoresearch Laboratories, West Grove, Pa). For the adhesion assays, calcium-dependent cell binding was tested using 10 μg/mL of a mouse E-selectin/human Fc chimera (R&D Systems, Minneapolis, Minn) in buffer containing 2 mM CaCl2 with or without 2 mM ethylene diamine tetracetic acid (EDTA) (a calcium ion chelator) or with 10 mM EDTA alone as a control. PE-conjugated F(ab′)2 goat antihuman IgG (Jackson Immunoresearch Laboratories) was used to detect bound E-selectin chimera. Data acquisition and analysis were performed using a FACSCanto instrument and BD FACSDiva software (Becton-Dickinson Biosciences, San Jose, Calif).
Cytokine Secretion Analysis
Cellular supernatants were collected after 72 hours of culture in serum-free media, and cytokine secretion was detected with the Human Cytokine Antibody Array 1 system (Ray Biotech, Inc., Norcross, Ga) using chemiluminescence. The data were quantified by densitometry using Image J software (National Institutes of Health, Bethesda, Md). Secreted interleukin 8 (IL-8) levels were analyzed with the human IL-8 enzyme-linked immunosorbent assay (ELISA) kit (Ray Biotech, Inc.) using a SpectraMax Plus 385 microplate reader at an optical density of 450 nm and SoftMax Pro 3.1.1 software (both from Molecular Devices, Sunnyvale, Calif).
Stimulation Assays With Neutrophils, Neutrophil-Conditioned Media, or Recombinant Human TNF-α
Neutrophils were isolated from peripheral blood of healthy human donors as described previously18 and with the approval of the Institutional Review Board (the Human Subjects Committee at the University of Minnesota). Cell viability was determined by Trypan blue exclusion. Neutrophil-conditioned media (NCM) was obtained from neutrophils (1.25 × 106 cells/mL) by culturing in Opti-MEM (Invitrogen) with penicillin/streptomycin for 2 hours. After serum-starvation of H1299 cells for 24 hours, the cells were cultured with neutrophils, NCM, or recombinant human TNF-α (rhTNF-α) for 48 hours. An optimal concentration of 5 ng/mL rhTNF-α was used for stimulation of sLeX expression on H1299 cells. Neutrophils were labeled with carboxyfluorescein succinimidyl ester (CFSE) fluorescent dye (Molecular Probes, Eugene, Ore) and then added to H1299 cells in the following neutrophil-to-cancer cell ratios: H1299 cells alone, 5:1 (1.25 × 106 neutrophils), 1:1 (2.5 × 105 neutrophils), and 1:5 (5 × 104 neutrophils), and an optimal 1:1 ratio was used in the experiments.
Cells were also cultured with a low-endotoxin azide-free (LEAF)-purified TNF-α–blocking antibody (10 μg/mL; BioLegend, San Diego, Calif) or an isotype-matched control MoAb (Invitrogen) for 48 hours. H1299 cells were analyzed for cell membrane expression of sLeX using CSLEX1 MoAb and for TNF-receptor 1 (TNF-R1) using an antihuman TNF-RI antibody (10 μg/mL; R&D Systems), and neutrophils were distinguished from cancer cells based on their morphology by forward and side scatter flow cytometric analysis. Changes in cellular morphology of cultured cells were assessed by light and fluorescent microscopic analysis at ×100 magnification.
RNA Isolation and Semiquantitative RT-PCR Analysis
FucT-III gene expression in H1299 cells that were either unstimulated or stimulated with 5 ng/mL rhTNF-α or NCM for 24 hours and 48 hours was verified with RT-PCR by amplification for 35 cycles as previously described.19 The FucT-III primers were 5′-CCT CCT GAT CCT GCT ATG GA-3′ (sense) and 5′-GCG GTA GGA CAT GGT GAG AT-3′ (antisense; (Genbank no. NM_000149), and the β-actin primers were 5′-GCT CGT CGT CGA CAA CGG CT-3′ (sense) and 5′-CAA ACA TGA TCT GGG TCA TCT TCT C-3′ (antisense; Genbank no. NM_001101).
Duplicate samples of 36 common human lung cancers, 4 normal lung tissues, and 8 inflamed lung tissues were obtained from the Tissue Procurement Facility at the University of Minnesota or from a human lung cancer tissue array (Pantomics, Inc., San Francisco, Calif). Tissues were used in accordance with the regulations of the University of Minnesota's Institutional Review Board. Sections were 4 μm thick, and they were fixed in 10% neutral buffered formalin and embedded in paraffin. Specimens were stained with hematoxylin and eosin by the University of Minnesota's Comparative Pathology Core Resource Facility.
Tissue sections were analyzed for the presence of the sLeX epitope using CSLEX1 MoAb and, for cytokeratins, using the mouse IgG1 MoAb Ab-3, clone Lu-5, which reacts with epithelial cells (Thermo Fisher Scientific, Fremont, Calif), as previously described.19 The staining procedure was modified to detect neutrophil elastase using a rabbit polyclonal primary antibody (0.5 μg/mL; Abcam, Cambridge, Mass). All antibody incubations were performed at 30°C. Pepsin Solution Digest-All 3 ready-to-use reagent (Invitrogen) was used for antigen retrieval. The secondary antibody was horseradish peroxidase-conjugated donkey antirabbit IgG (BioFX Laboratories, Owings Mills, Md) at 1:500 dilution.
Histologic Grading System
Lung tissues were evaluated semiquantitatively for reactivity with CSLEX-1 MoAb to indicate sLeX-positive cells and quantitatively for reactivity with the neutrophil elastase antibody to determine the number of neutrophils present in tissues. For individual lung tissues, sections were evaluated in 5 random fields at ×100 magnification. For the tissue array, the entire tissue core was examined. Positive staining was defined as cells in which cytoplasmic and/or cell membrane labeling with the relevant antibody was detected. For both antibodies, the presence of positive staining was scored from 0 to 3. For CSLEX-1 MoAb reactivity, a score of 0 indicated no visible immunostaining, a score of 1 indicated that <25% of the total tissue area was positive, a score of 2 indicated that 26% to 50% of the total tissue area was positive, and a score of 3 indicated that >50% of the total tissue area was positive. For reactivity with the neutrophil elastase antibody, a score of 0 indicated that cells were not visibly immunostained, a score of 1 indicated that 1 to 10 cells in the total tissue area were positive, a score of 2 indicated that 11 to 35 cells in the total tissue area were positive, and a score of 3 indicated that >35 cells in the total tissue area were positive.
Twenty-four-well BioCoat Matrigel Invasion Chambers (Becton-Dickinson, Bedford, Mass) were used for the tumor cell invasion assays according to the manufacturer's instructions. Chambers were cultured for 24 hours under standard conditions. The invading cells were stained and counted in 5 random fields at ×100 magnification. Digital images were captured using a Zeiss Axiovert 200 inverted microscope and an AxioCam MRc camera with AxioVision 4.1 software (Carl Zeiss Inc., Oberkochen, Germany).
Differences between groups were analyzed using a 2-tailed Student t test, with 2-sample unequal variance where appropriate. Reported P values were considered significant at P ≤ .05.
Characteristics of Transfected H1299 Cells and Response to rhTNF-α
Stably transfected H1299 and A549 cells, but not mock-transfected cells, highly expressed FucT-III mRNA by RT-PCR and were positive for sLeX by flow cytometry using the CSLEX-1 MoAb. Data are shown only for transfected H1299 cells that were 57% positive for sLeX compared with mock-transfected cells (P < .01) (Fig. 1A,B). Cell viability after transfection was 98% to 99%. We investigated the effects of TNF-α on sLeX expression because it is 1 of the major cytokines produced in a tumor microenvironment. Cells were stimulated with rhTNF-α at concentrations between 1 and 100 ng/mL for 24 and 48 hours, and we observed that exposure to 5 ng/mL rhTNF-α for 48 hours was the lowest dose that produced significant up-regulation of sLeX expression on FucT-III–transfected H1299 cells (74%) compared with mock-transfected cells (P < .01), and expression was greater than that in unstimulated transfected cells (74% vs 57%; P < .05) (Fig. 1B). All groups of cells expressed similar levels of TNF-R1 (57%-63%; data not shown).
The E-selectin binding ability of mock-transfected and FucT-III–transfected H1299 and A549 cells that expressed the sLeX epitope and were either unstimulated or stimulated with 5 ng/mL rhTNF-α was investigated by flow cytometry using a mouse E-selectin chimera. Mock-transfected H1299 cells did not appreciably bind to E-selectin regardless of whether or not they were stimulated with rhTNF-α, whereas unstimulated, FucT-III–transfected H1299 cells bound to E-selectin in a calcium-dependent manner (80%; P < .01) (Fig. 1C). After stimulation with rhTNF-α, E-selectin binding of transfected H1299 cells was increased significantly more (92%; P < .05). Appreciable binding did not occur in the presence of EDTA, as expected. Similar results were obtained for FucT-III–transfected and mock-transfected A549 cells (data not shown). These results indicate that NSCLC cells that express sLeX, but not cells that lack sLeX, specifically bind to E-selectin and that binding is responsive to stimulation with TNF-α.
To investigate whether FucT-III gene expression influenced characteristics that might have an impact on invasion and metastasis, we compared the cytokine expression profile of FucT-III–transfected H1299 cells with that of unmanipulated cells after 72 hours in culture using a human anticytokine antibody membrane array. Transfected cells secreted increased levels of growth-regulated oncogene-α (GRO-α), IL-8, and mast cell proteinase-1 (MCP-1); but levels of granulocyte-macrophage–colony-stimulating factor (GM-CSF) and MCP-3 were down-regulated (P < .05) (Fig. 2A). Because IL-8 is a potent neutrophil chemoattractant, we quantified IL-8 secretion using a human IL-8 ELISA kit. FucT-III–transfected cells secreted 2-fold higher levels of IL-8 than mock-transfected cells (P < .05) (Fig. 2B). Our results suggest that sLeX expression on NSCLC cells is associated with up-regulated secretion of neutrophil chemoattractants.
FucT-III–Transfected H1299 Cells Cultured With Neutrophils or With NCM Up-Regulate sLeX
We investigated whether neutrophils regulated sLeX expression on H1299 cells. Serum-starved cells were cultured for 48 hours with neutrophils or with NCM, and the effects on sLeX expression were determined by flow cytometry. After culture of neutrophils with FucT-III–transfected cells at ratios of 5:1 and 1:1, the latter cells were 81% and 68% positive for sLeX, respectively, compared with unstimulated FucT-III–transfected cells (36% sLeX-positive; P < .01) (Fig. 3A). After stimulation with NCM, 24% of FucT-III–transfected cells were sLeX-positive compared with unstimulated FucT-III–transfected cells (13% sLeX-positive) (Fig. 3B). Our results suggested that neutrophils and their secreted factors stimulated sLeX expression on NSCLC cells.
Up-Regulation of sLeX by Neutrophils and NCM Is TNF-α Dependent
We treated FucT-III–transfected H1299 cells with an anti-TNF-α–blocking MoAb or an isotype-matched control MoAb (10 μg/mL) in the presence of a 1:1 ratio of neutrophils, NCM, or 5 ng/mL rhTNF-α for 48 hours to determine whether TNF-α could be responsible for the up-regulation of sLeX. SLeX expression was significantly increased on control MoAb-treated, FucT-III–transfected cells when cultured with neutrophils or NCM (P < .05) or with rhTNF-α (P < .01) compared with cells that were exposed to control MoAb alone (Fig. 3C). In the presence of an anti-TNF-α–blocking MoAb, sLeX expression was significantly down-regulated on FucT-III–transfected cells that were cultured with neutrophils, NCM, or rhTNF-α compared with similarly cultured cells that were treated with the control MoAb (P < .05). Antibody treatment did not affect basal sLeX expression. Our findings suggest that neutrophil-derived TNF-α stimulates sLeX production on NSCLC cells that express the FucT-III gene.
Steady-State FucT-III mRNA Levels Are Not Altered by rhTNF-α or NCM
To determine whether the up-regulation of sLeX that we observed after stimulation of FucT-III–transfected H1299 cells with rhTNF-α or NCM was because of increased steady-state mRNA levels of the FucT-III gene, we compared FucT-III mRNA levels in unstimulated cells and cells that were stimulated with 5 ng/mL rhTNF-α or NCM for 24 or 48 hours by RT-PCR. We did not examine mRNA from cocultures of H1299 cells and neutrophils, because they were indistinguishable. FucT-III mRNA levels were similar between unstimulated cells and cells that were stimulated with rhTNF-α or NCM for 24 or 48 hours. The 48-hour time point is illustrated in Figure 3D. Although not conclusive, the results suggest that molecular events downstream from the synthesis of primary transcript mRNA may be regulated by neutrophil-derived TNF-α.
Neutrophils, NCM, and rhTNF-α Induce Morphologic Changes in H1299 Cells
We assessed whether FucT-III–transfected H1299 cells or mock-transfected cells that were cocultured with neutrophils, NCM, or rhTNF-α for 48 hours underwent morphologic changes indicative of a more metastatic phenotype. At ×100 magnification, the morphology of unstimulated FucT-III–transfected H1299 cells was not altered from that of unstimulated, mock-transfected cells (Fig. 4A, B) and was similar to the morphology of a 1:1 ratio of neutrophils cocultured with mock-transfected H1299 cells in the absence or presence of 10 μg/mL anti-TNF-α–blocking MoAb (Fig. 4C,E). FucT-III–transfected H1299 cells cocultured with neutrophils formed viable, nonadherent aggregates of elongated cells (Fig. 4D). When cocultures were exposed to the anti-TNF-α–blocking MoAb, the cancer cells reverted to their original adherent morphology (Fig. 4F). In cocultures, CFSE-labeled neutrophils were viable but distinct from FucT-III–transfected cell aggregates when examined at ×100 magnification under brightfield (Fig. 4G) and fluorescent (Fig. 4H) microscopy. In the absence of cancer cells, most neutrophils died in culture after 48 hours (data not shown).
After 48 hours in culture, we did not observe morphologic differences between the unstimulated, FucT-III–transfected H1299 cells that were treated with the isotype control MoAb and the cells that were treated with the anti-TNF-α–blocking MoAb (10 μg/mL), (Fig. 5A,B). Control MoAb-treated, FucT-III–transfected cells cultured with 5 ng/mL rhTNF-α formed an adherent monolayer of elongated cells, and similarly treated cells cultured with NCM grew as adherent, elongated colonies (Fig. 5C,E). In contrast, FucT-III–transfected cells that were treated with the anti-TNF-α–blocking MoAb and cultured with 5 ng/mL rhTNF-α or with NCM grew as a monolayer with an appearance similar to that of unstimulated cells (Fig. 5B,D,F). Our results suggest that TNF-α produced by neutrophils caused the altered morphology of cancer cells.
FucT-III Regulates H1299 Cell Invasiveness, Which Is Enhanced by rhTNF-α Stimulation
To determine whether the morphologic changes that we observed in H1299 cells were indicative of alterations in their metastatic behavior, we examined the invasiveness of cells, which is a key metastatic property, using in vitro Matrigel invasion assays. Unstimulated, FucT-III–transfected H1299 cells were more invasive than unstimulated mock-transfected cells (P < .05) (Fig. 6A,C,E). After stimulation with 5 ng/mL rhTNF-α for 24 hours, FucT-III–transfected cells, but not mock-transfected cells, were more invasive than the corresponding unstimulated cells (P < .01) (Fig. 6A,C-F). FucT-III–transfected cells that were stimulated with NCM for 24 hours were more invasive than unstimulated FucT-III transfected cells (Fig. 6B,E,G) (P < .01). Cells that were treated with the anti-TNF-α–blocking MoAb and cultured with NCM were less invasive than the cells that were cultured with NCM alone (Fig. 6B,G,H) (P < .01). The results indicate that sLeX expression on H1299 cells enhanced their invasiveness, which was augmented further by stimulation with TNF-α.
Neutrophils Colocalize With sLeX-Positive Lung Adenocarcinoma and Squamous Cell Carcinoma Cells
We examined 36 duplicate patient samples of human lung cancers and 4 normal lung tissues for sLeX expression and neutrophil infiltration by immunohistochemistry. Of these malignant lung tumors, 12 were adenocarcinomas, and 12 were squamous cell carcinomas. The remaining 12 malignant tumors consisted of 3 adenosquamous carcinomas, 3 small cell carcinomas, 1 each of a bronchioalveolar and a large cell carcinoma, and 4 papillary carcinomas. We also examined an additional 8 inflamed tissues. We observed that 78% of the malignant lung cancers were reactive with CSLEX1 MoAb, indicating that these tissues were sLeX-positive (sLeX staining scores of 1-3). Of these malignant tumors, 61% were tumors in which >25% of the total tissue area examined was positive for sLeX (staining scores of 2-3). Higher numbers of neutrophils were stained with neutrophil elastase antibody (neutrophil scores of 2-3) in 71% of highly sLeX-positive tumor tissue samples with scores of 2 or 3. All of the inflamed tissues contained high numbers of stained neutrophils, and 63% were sLeX-positive.
We focused on adenocarcinomas and squamous cell carcinomas, because they are common tumor types. All 12 adenocarcinomas were sLeX-positive, and 83% of those sLeX-positive tumors were infiltrated with neutrophils. All 12 squamous cell carcinomas contained neutrophils, and 50% had high neutrophil scores of 2 or 3. Of those squamous cell carcinomas, 83% were sLeX-positive. In contrast, all 4 normal lung tissues were sLeX-negative but contained variable numbers of neutrophils. Although our sample size was too small to draw statistical conclusions, we observed that, in 75% of adenocarcinoma samples and in 58% of squamous cell carcinoma samples, areas of positive sLeX expression were similar to areas of neutrophil infiltration. Representative examples of these results are provided in Figure 7.
Despite recent advances made in earlier diagnosis and treatment, NSCLC continues to have a poor prognosis and survival, and these outcomes are commonly because of locally invasive or metastatic disease.1 We expect that the identification of novel molecular and cellular determinants that contribute to disease progression of NSCLC will lead to the development of strategies to inhibit or limit cancer growth, invasion, and metastasis and improve therapeutic efficacy. We investigated the influence of neutrophils and TNF-α on regulation of sLeX expression in NSCLC and the importance of these interactions in promoting metastatic properties of tumor cells.
The cytokine TNF-α is produced predominantly by macrophages and T-cells but also by neutrophils, and it regulates diverse cellular functions, including inflammation and immunity, homeostasis, and apoptosis. TNF-α also participates in cell-mediated killing and necrosis of certain tumors.20 Recent evidence suggests that tumor-associated neutrophils promote a metastatic phenotype through secretion of TNF-α.14, 15 TNF-α regulates expression of sLeX epitopes on human lung carcinomas.17, 21 NSCLCs with high expression of sLeX have increased metastatic potential.4 By using NSCLC cell lines, in the current study, we demonstrated that FucT-III gene expression resulted in the production of cell membrane sLeX, the release of neutrophil chemoattractants, and increased invasiveness of tumor cells. Neutrophil-derived TNF-α stimulated sLeX expression on tumor cells, caused them to be nonadherent, and further enhanced their invasiveness. These changes were indicative of increased metastatic behavior of NSCLC cells. In human lung specimens of NSCLC, we observed that neutrophils were recruited to tumor areas of sLeX expression. Thus, carbohydrate-mediated interactions between NSCLC cells and neutrophils may facilitate invasion and possibly metastasis of tumor cells.
In an in vivo rat model, it was demonstrated that tumor-elicited neutrophils, but not normal neutrophils, enhanced the metastatic potential and invasiveness of tumor cells.11 Currently, we are testing the potential mechanisms involving sLeX and neutrophils that may be responsible for the increased invasiveness of NSCLC cells. It is well established that invasion and metastasis of malignant cells require degradation of basement membranes and of the extracellular matrix involving the matrix metalloproteinases MMP-2 and MMP-9. Increased levels of MMP-2 and MMP-9 and of sLeX in NSCLC have been correlated with an advanced stage of disease, distant metastasis, and poor survival.4, 22, 23 sLeX is synthesized by members of the family of α(2,3)-sialyltransferases and α(1,3/4)-fucosyltransferases (FucTs),24, 25FucT-III is up-regulated in NSCLC and determines the amount of sLeX antigens in lung cancer tissues.26 We speculate that, in NSCLC, the FucT-III enzyme may glycosylate and up-regulate the expression or activity of MMP-2, MMP-9, or other matrix metalloproteinases, resulting in more invasive behavior of tumor cells. In our current investigation, steady-state FucT-III mRNA levels were unchanged after the stimulation of FucT-III–transfected H1299 cells with TNF-α or NCM, although sLeX expression was increased. It is conceivable that post-transcriptional modifications of FucT-III mRNA in these cells may have occurred after exposure to TNF-α or NCM, such as alterations in mRNA stability or in mRNA processing and transport. Alternatively, FucT-III enzyme function may be changed by post-translational modifications of the protein, such as glycosylation or phosphorylation in response to TNF-α stimulation, which may result in up-regulation of sLeX. These possibilities require further investigation.
Many cancers are able to recruit neutrophils that release substances to modify tumor growth and invasiveness.27 The presence of tumor-associated neutrophils in bronchioloalveolar carcinoma has been linked to poor clinical outcomes, including decreased survival, and these tumor cells produce IL-8, which is a chemoattractant for neutrophils.10 In agreement with those studies, we demonstrated that H1299 FucT-III–transfected cells expressing sLeX produced significantly increased levels of the neutrophil chemoattractants GRO-α, MCP-1, and IL-8. In addition, sLeX expression and invasiveness of these lung carcinoma cells was augmented by neutrophil-derived TNF-α. Thus, our study indicates that a novel mechanism of tumor progression of NSCLC exists whereby TNF-α secreted by neutrophils present in NSCLC may up-regulate FucT-III gene expression and, thus, sLeX on tumor cells that promotes invasiveness. Tumor cells that express sLeX secrete higher amounts of IL-8, which may recruit neutrophils to tumors to continue driving the invasive process, ultimately resulting in metastasis. We observed significant down-regulation of GM-CSF and MCP-3 in H1299-transfected cells. These chemokines promote eosinophil recruitment to tumors, which may also facilitate NSCLC progression.28
In summary, our findings point to a unique role of neutrophils producing TNF-α in NSCLC tumor progression: enhancement of sLeX expression on NSCLC cells, alterations in morphology consistent with increased invasiveness, and, together with sLeX, induction of increased invasive behavior of NSCLC cells. These data support the identification of a novel mechanism of lung tumor progression that may be manipulated to inhibit the invasiveness and metastasis of lung cancers. Thus, our findings lay the groundwork for a promising new approach to lung cancer therapy.
We thank the following resources at the University of Minnesota's Masonic Cancer Center: members of the Comparative Pathology Shared Resource core facility for assistance with immunohistochemistry and the Tumor Biology and Progression Research Program for discussion of the data.
This work was supported in part by funds from the National Institutes of Health (National Cancer Institute grant 5KO8CA111829-04).