Cancer Cell Biology
Ovarian cancers overexpress the antimicrobial protein hCAP-18 and its derivative LL-37 increases ovarian cancer cell proliferation and invasion
Version of Record online: 24 OCT 2007
Copyright © 2007 Wiley-Liss, Inc.
International Journal of Cancer
Volume 122, Issue 5, pages 1030–1039, 1 March 2008
How to Cite
Coffelt, S. B., Waterman, R. S., Florez, L., Bentrup, K. H. z., Zwezdaryk, K. J., Tomchuck, S. L., LaMarca, H. L., Danka, E. S., Morris, C. A. and Scandurro, A. B. (2008), Ovarian cancers overexpress the antimicrobial protein hCAP-18 and its derivative LL-37 increases ovarian cancer cell proliferation and invasion. Int. J. Cancer, 122: 1030–1039. doi: 10.1002/ijc.23186
- Issue online: 24 DEC 2007
- Version of Record online: 24 OCT 2007
- Manuscript Accepted: 24 AUG 2007
- Manuscript Received: 18 APR 2007
- NIH. Grant Numbers: AI056229, 1P20RR20152-01
- Cancer Association of Greater New Orleans (CAGNO)
- ovarian cancer;
The role of the pro-inflammatory peptide, LL-37, and its pro-form, human cationic antimicrobial protein 18 (hCAP-18), in cancer development and progression is poorly understood. In damaged and inflamed tissue, LL-37 functions as a chemoattractant, mitogen and pro-angiogenic factor suggesting that the peptide may potentiate tumor progression. The aim of this study was to characterize the distribution of hCAP-18/LL-37 in normal and cancerous ovarian tissue and to examine the effects of LL-37 on ovarian cancer cells. Expression of hCAP-18/LL-37 was localized to immune and granulosa cells of normal ovarian tissue. By contrast, ovarian tumors displayed significantly higher levels of hCAP-18/LL-37 where expression was observed in tumor and stromal cells. Protein expression was statistically compared to the degree of immune cell infiltration and microvessel density in epithelial-derived ovarian tumors and a significant correlation was observed for both. It was demonstrated that ovarian tumor tissue lysates and ovarian cancer cell lines express hCAP-18/LL-37. Treatment of ovarian cancer cell lines with recombinant LL-37 stimulated proliferation, chemotaxis, invasion and matrix metalloproteinase expression. These data demonstrate for the first time that hCAP-18/LL-37 is significantly overexpressed in ovarian tumors and suggest LL-37 may contribute to ovarian tumorigenesis through direct stimulation of tumor cells, initiation of angiogenesis and recruitment of immune cells. These data provide further evidence of the existing relationship between pro-inflammatory molecules and ovarian cancer progression. © 2007 Wiley-Liss, Inc.
The contribution of chronic infection and inflammation to tumorigenesis and tumor progression has recently received a great deal of attention. It has been shown that many cancers display a strong association between leukocyte infiltration and disease progression.1 Among these, ovarian cancer remains one of the most prominent tumor types. Epidemiological reports have revealed that women exposed to pro-inflammatory agents are at increased risk for developing ovarian cancer whereas down-regulating inflammatory events may prevent ovarian tumorigenesis.2, 3 Other studies demonstrate ovarian cancer's induction of tolerance through recruitment of immunosuppressive leukocytes and the importance of many pro-inflammatory soluble factors, such as tumor necrosis factor α, has been established.4, 5 Although advances have been made in the understanding of ovarian tumor development and progression, the involvement of a number of key pro-inflammatory molecules has yet to be examined including host-derived antimicrobial peptides.
The role of one such peptide, termed human cationic antimicrobial protein 18 (hCAP-18/LL-37), in wound healing and inflammation has been well characterized.6, 7, 8, 9, 10, 11 Leukocytes and epithelial cells synthesize hCAP-18 as a precursor protein (referred to here as prepro-LL-37) consisting of 3 domains: a signal peptide, cathelin domain and LL-37 peptide. Once localized to subcellular compartments, the signal peptide is enzymatically removed (hCAP-18 or pro-LL-37). Leukocytes and epithelial cells are prompted to release increased amounts of hCAP-18/LL-37 into the extracellular space after tissue injury or other pro-inflammatory events making hCAP-18/LL-37 susceptible to serine proteases.8, 9, 10, 12, 13 The resulting cleavage of hCAP-18/LL-37 yields 2 independent and functional proteins.14, 15, 16 The N-terminal cathelin pro-domain serves as a bacterial protease inhibitor, while the 37 amino acid C-terminal peptide, LL-37, and exhibits a broad range of antimicrobial and wound healing activities. Some of these activities include direct eradication of pathogens, stimulation of angiogenesis, reepithelialization of healing skin and the induction of other pro- and antiinflammatory chemokines or chemokine receptors.6, 8, 17, 18, 19, 20, 21 In addition, LL-37 acts as a potent chemoattractant for monocytes/macrophages, mast cells, neutrophils and CD4+ T cells. LL-37 elicits its chemotactic effects—with the exception of mast cells—through formyl peptide receptor like-1 (FPRL-1), a G-protein coupled receptor.22, 23, 24, 25, 26 On nonimmune cells, LL-37 treatment has also been shown to transactivate the epidermal growth factor (EGF) receptor to stimulate proliferation as well as chemotaxis.11, 19, 27
LL-37's association with inflammation and angiogenesis implicates a role for this peptide in tumorigenesis. However, knowledge pertaining to hCAP-18/LL-37 expression and function in human cancers is limited. Thus, the goal of this study was to characterize hCAP-18/LL-37 expression in cancers of the ovary and to examine the effects of LL-37 on ovarian cancer cells. We report here that hCAP-18/LL-37 is significantly overexpressed in ovarian cancers when compared to normal ovarian tissue. Both the bioactive LL-37 peptide and its pro-form, hCAP-18, were identified from lysates of ovarian tumors. In addition, recombinant LL-37 peptide stimulated ovarian cancer cell proliferation, migration, invasion and matrix metalloproteinase (MMP) secretion. The data presented indicate that hCAP-18/LL-37 may be critical for human ovarian cancer progression and as such, may represent a potential therapeutic target.
Material and methods
Tumor tissue lysate array
DiscoverLight™ Tissue Lysate arrays (Pierce Biotechnology, Rockford, IL) were blocked with 5% milk protein in phosphate buffered saline containing 0.2% Tween-20 (PBST), then incubated with an anti-hCAP-18/LL-37 monoclonal antibody (1:500, clone 1-1C12; Hycult Biotechnology, Uden, The Netherlands) overnight at 4°C followed by horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG antibodies (1:5,000; Amersham Biosciences, Pittsburgh, PA). Enhanced chemiluminescence system (ECL; Amersham Biosciences) was added following washing steps and the membrane was exposed to X-ray film (Kodak, Rochester, NY).
Human ovarian tissue arrays contained 144 sections (BioChain Institute, Hayward, CA and US Biomax, Rockville, MD): 40 were normal tissue and 104 were tumor tissue. Epithelial-derived ovarian tumors comprised 88 out of the 104 total tumor specimens. The tissue sections were deparaffinized before boiling in 10 mM sodium citrate, pH 6.0. Sections were incubated with 3% hydrogen peroxide before blocking (10% human serum, 10% goat serum in stain buffer [BD Biosciences, San Jose, CA]). Immunostaining was performed using monoclonal anti-hCAP-18/LL-37 (1:50, clone 3D11; Hycult Biotechnology), anti-CD45 (1:50; DAKO, Carpinteria, CA) and anti-CD31 (1:100; Chemicon, Temecula, CA) antibodies overnight at 4°C. This was followed by incubation with Envision + Dual Link (DAKO) or antibodies provided in Chemicon's Blood Vessel Staining Kit following the manufacturer's instructions. Signals were detected using diaminobenzidine as substrate (DAKO). Omission of primary antibodies served as a negative control. Sections were counterstained with hematoxylin, dehydrated, mounted and coverslipped for bright field viewing using a Zeiss Axioplan 2 fluorescence microscope with Intelligent Imaging Innovations software (SlideBook ver. 4). Information regarding histopathology and International Federation of Gynecology and Obstetrics (FIGO) staging were obtained from the companies. The coauthor pathologist (L.F.) reviewed and confirmed all samples with respect to histopathology and cell type-specific staining.
For evaluation of the staining, 3 investigators examined the slides in a blinded manner. Each section was given an intensity rating (0–3) and a percentage of cells positive rating (0 = 0%, 1= 1–19%, 2 = 20–79%, 3 = 80–100%). An overall score was calculated by multiplying positivity with staining intensity. The average number of blood vessels was taken from 3 400× fields of view for each section.
Western blot analysis
Frozen primary ovarian tumors were obtained from the National Disease Research Institute then pulverized in liquid nitrogen using a cooled mortar and pestle. T-PER (Pierce) protein extraction reagent containing protease inhibitors was added to ground tumor samples, incubated on ice for 30 min, and sonicated for 30 sec. For ovarian cancer cell lines, whole cell lysates were isolated using M-PER (Pierce) lysis buffer containing protease inhibitors following the manufacturer's instructions. Protein concentration was determined using the BCA™ Protein Assay kit (Pierce). A 10–20% SDS-polyacrylamide gel was loaded with 50 μg of protein or 20 ng recombinant LL-37 (Innovagen, Lund, Sweden) for separation then transferred to nitrocellulose. The remainder of the procedure was carried out as previously described using anti-hCAP-18/LL-37 (1:250; Phoenix Pharmaceuticals, Belmont, CA) polyclonal antibodies.28 β-actin (1:5,000; Sigma, St. Louis, MO) was used to ensure equal loading.
Hycult's LL-37 ELISA kit was used to measure hCAP-18/LL-37 in ovarian tissue lysates according to the manufacturer's instructions.
Ovarian cancer cell lines Hs832.Tc (benign ovarian cyst), OV-90 (papillary serous adenocarcinoma), SK-OV-3 (adenocarcinoma), TOV-112D (endometrioid adenocarcinoma), TOV-21G (clear cell adenocarcinoma) were obtained from the American Type Culture Collection and propagated according to their recommendations. HEY (xenograft HX-62, papillary cystadenocarcinoma) and OVCAR-3 (adenocarcinoma) cell lines were maintained in RPMI-1640 (Gibco, Carlsbad, CA) containing 10% fetal bovine serum (FBS; Atlanta Biologicals, Norcross, GA) and 100 units/mL pencillin and 100 mg/mL streptomycin (Gibco).
Ovarian cancer cells were seeded in 96-well plates (2 × 103 HEY cells and 5 × 103 OV-90 or SK-OV-3 cells per well) and allowed to adhere overnight. The next day, cells were washed 3× with PBS and serum-free medium was added. After 24 hr, cells were treated with LL-37 as indicated or 10 ng/mL EGF (R&D Systems, Minneapolis, MN) in medium containing 5% fetal bovine serum. Each experimental condition was tested in replicates of 8. Cells were allowed to proliferate for 48 hr. The amount of DNA in each well was quantified using Invitrogen's CyQuant® NF Cell Proliferation Kit following the manufacturer's instructions. Relative fluorescence units were obtained with a FLUOstar Optima microtiter plate reader (BMG Labtech, Durham, NC). Results are presented as the mean ± standard error of the mean (s.e.m.) of 3 or more experiments.
Boyden chamber migration assay
Chemoattractants in medium containing 0.5% bovine serum albumin (BSA) or 0.5% fetal bovine serum (FBS) were added to the lower compartment of a 48-well modified Boyden chamber in triplicate and overlaid with a porous membrane (8 μm; Neuro Probe, Gaithersburg, MD) that was precoated with 0.1 mg/mL gelatin (Bio-Rad Laboratories, Hercules, CA). LL-37 was added to a final concentration of 0.1, 1 or 10 μg/mL; EGF was used at a concentration of 10 ng/mL. The chamber was assembled and cells, serum-starved for 24 hr, were added to the upper chamber at a density of 2 × 105 cells per well. The apparatus was incubated at 37°C for 6hr to allow for cell migration. After that period, the apparatus was disassembled and cells were stained using Diff-Quick Stain Set (Dade Behring, Deerfield, IL). Cells on the top of the filter, which had not migrated through, were wiped off. Migration was quantified by counting the nuclei that passed through the filter from a minimum of 6 fields of view (400×) for 3 replicates. Images were collected using a Nikon TE300 inverted microscope (DP Controller v126.96.36.199, Olympus Optical Company, TX). Experiments were repeated a minimum of 3 times. Results are expressed as fold increase over untreated, control cells.
FluoroBlok™ (BD Biosciences) inserts (8 μm) were coated with 50 μL of growth factor-reduced Matrigel™ (BD Biosciences) diluted 1:10 in serum-free medium and allowed to polymerize overnight at 37°C. The next day chemoattractants in medium containing 5% fetal bovine serum were added to the bottom of a 24-well plate in duplicate and inserts were placed in the wells. LL-37 was added to a final concentration of 0.1, 1 or 10 μg/mL; EGF was used at a concentration of 10 ng/mL. Serum-starved cells were seeded onto each Matrigel™-coated insert at a density of 1.5 × 105 cells per insert. After invasion was allowed to occur for 16 hr at 37°C, the inserts were placed into a new well containing 2 μg/mL Calcein AM (Molecular Probes) in Hank's balanced salt solution (Gibco). Inserts were incubated for an additional hour at 37°C to fluorescently label the invaded cells. Images were taken using a Nikon TE300 inverted epifluorescence microscope and relative fluorescent units were obtained with a fluorescent plate reader (BMG Labtech). Experiments were repeated a minimum of 3 times. Results are expressed as fold increase over untreated, control cells.
HEY and SK-OV-3 cells at a density of 1 × 105 were plated in 24-well plates and allowed to adhere overnight. After washing 3× with PBS, serum-free medium was added to the cells for 24hr. Growth factors (5 μg/mL LL-37 and 10 ng/mL EGF) were mixed with medium containing 0.5% fetal bovine serum and then added to the cells. Conditioned medium was collected after 24 hr and total protein concentrations were determined using Bio-Rad's protein assay. Protein (4 μg) was loaded in 10% polyacrylamide gels containing 0.1% gelatin or 12% polyacrylamide gels containing 0.05% casein (Invitrogen). Electrophoresis was performed as recommended by the manufacturer then gels were incubated with the following: renaturing buffer, 30 min, room temp.; developing buffer, 30 min, room temp.; and developing buffer, overnight, 37°C. The next day, gels were washed 3× with water for 5 min each. SimplyBlue™ SafeStain (Invitrogen) was added to gels for 1 hr then washed with water for 1 hr. Gels were incubated with Gel-Dry™ Drying Solution (Invitrogen) for 5 min before being placed between 2 pieces of cellophane. After gels dried overnight, zymogen activation was recorded using the Alpha Innotech (San Leandro, CA) gel documentation system. The experiment was performed at least 3 times.
HEY and SK-OV-3 ovarian cancer cell lines were treated with 5 μg/mL LL-37 and 10 ng/mL EGF for 6 hr. Total RNA was isolated using the RNeasy® Mini Kit (Qiagen, Valencia, CA). Purified RNA was treated with TURBO DNA-free™ (Ambion, Foster City, CA) followed by reverse transcription using 1 μg of RNA and the iScript™ cDNA Synthesis kit (Bio-Rad). For real-time PCR analysis, 50 ng of cDNA was amplified in 1X iQ™ SYBER Green Supermix (Bio-Rad) with 300 nM of MMP-2, -9, -14, or uPA specific primer sets (Integrated DNA Technologies, Coralville, IA; sequences available upon request) using the iCycler iQ5 Real-Time PCR Detection System (Bio-Rad). As a reference gene, primers for 36B4 (sequences available upon request) at a concentration of 200 nM were used. Experimental samples were run on the same plate as the internal control (36B4) in triplicate. Real-time PCR was performed 4 times from 4 independent RNA preparations. PCR parameters were as follows: 95°C for 10 min followed by 40 cycles at 95°C for 15 sec and 60°C for 1 min. After amplification was complete, melt curves were generated to confirm specificity of each primer pair beginning with 65°C and increasing 0.5°C for 61 cycles. Differences in gene expression were determined by the Quantitative Comparative CT (threshold value) method.
Data from treated groups were compared to untreated groups and significant differences were determined by the Student's t-test (paired and unpaired), Spearman's rank correlation and one-way analysis of variance (ANOVA) followed by Dunnett post hoc tests using GraphPad Prism software as appropriate. A p value < 0.05 was considered statistically significant.
hCAP-18/LL-37 expression is upregulated in several human tumors
The expression of hCAP-18/LL-37 in normal and cancerous tissue was first examined using a human tumor tissue protein lysate array. In this array, normal and tumor protein lysates taken from the same patient were spotted in duplicate on nitrocellulose (refer to www.pierce.net for a complete list of arrayed tissues). Ovarian, bladder and thyroid cancer lysates exhibited a pronounced upregulation of hCAP-18/LL-37 expression when compared to their matched, normal controls (Fig. 1). Previously published observations have shown by immunohistochemistry that hCAP-18/LL-37 expression is decreased in colonic cancer tissue when compared to normal colonic tissue.29 Our approach detected a negligible difference in expression between normal colonic and cancer tissue. Taken together, these results suggest that increased hCAP-18/LL-37 expression is not a general phenomenon of tumor tissue and that increased hCAP-18/LL-37 expression may be critical to tumorigenesis of specific tissues. Thus, these data prompted a more focused analysis of the pro-inflammatory hCAP-18/LL-37 in ovarian tissue because of the causal relationship between inflammation and ovarian cancer.2, 3, 30
hCAP-18/LL-37 is highly expressed in ovarian tumors when compared to normal ovary
Immunohistochemistry was performed on commercially available human biopsy arrays containing both cancerous and noncancerous ovarian tissue using an anti-hCAP-18/LL-37 monoclonal antibody. In normal ovarian tissue (n = 40), staining was most prominent in stromal cells (Figs. 2a–2d). No staining was observed in squamous ovarian surface epithelium (Figs. 2a and 2b; black arrows). Cells lining primordial follicles as well as granulosa cells of mature follicles exhibited positive staining for hCAP-18/LL-37 (Figs. 2b–2d; orange arrows). Staining was also noted in endothelium (Figs. 2c and 2d; yellow arrows). No positive staining was detected in the absence of the primary antibody (data not shown).
The biopsy arrays contained 104 unique ovarian tumors of various stages, grades and histological subtypes including serous, mucinous, endometrioid, transitional cell, clear cell, thecoma, granulosa cell, signet-ring cell, teratoma and dysgerminoma. Of the 104 tumor sections, 88 were epithelial-derived ovarian tumors. Immunohistochemical analysis confirmed that all ovarian tumors expressed varying levels of hCAP-18/LL-37 (Figs. 2e–2h). Regardless of subtype, tumor cells were the predominant hCAP-18/LL-37-expressing cell type; although, expression was also noted in stromal cells such as endothelial cells, leukocytes and fibroblasts (Figs. 2e–2h; yellow arrows). To determine the relative hCAP-18/LL-37 expression differences between normal and cancerous ovarian tissue, staining was measured as described in the Materials and methods Section and used in statistical analysis. Not surprisingly, ovarian tumor tissue expressed significantly more hCAP-18/LL-37 than normal ovarian tissue (p < 0.001; Fig. 2i).
Increased hCAP-18/LL-37 expression correlates with immune cell infiltration, angiogenesis and colocalizes with CD45 and CD31 in ovarian tumors
Confirmation of hCAP-18/LL-37 expression in ovarian tumor-associated leukocytes was carried out using the established immune cell marker CD45. As expected, hCAP-18/LL-37 and CD45 staining colocalized in tumor stroma (Fig. 3a). From this observation, it was hypothesized that tumor-derived LL-37 may contribute to the recruitment of immune cells because of its known chemotactic role in wound healing and inflammation.11, 23, 25, 26, 27, 31 To explore this notion, we decided to focus on epithelial-derived ovarian tumors. CD45 staining in epithelial-derived tumors (n = 88) was scored in the same manner as hCAP-18/LL-37 staining and these values were used in statistical analysis. Among all the epithelial-derived ovarian tumors examined, a significant positive correlation (p = 0.043, r = 0.216) was observed between CD45 and hCAP-18/LL-37 intensity scores (Fig. 3b).
Endothelial cells were another stromal cell type where hCAP-18/LL-37 was detected. CD31 expression was used to validate this observation. Colocalization of hCAP-18/LL-37 and CD31 was evident in most blood vessels (Fig. 3c). Previous reports have shown the peptide LL-37 to be a potent angiogenic factor.6, 17 Therefore, increased hCAP-18/LL-37 protein expression in ovarian tumors may be associated with increased angiogenesis. In epithelial-derived ovarian tumors, microvessel density was measured as described in the Materials and methods Section (n = 88). A significant positive correlation (p = 0.046, r = 0.214) was observed between hCAP-18/LL-37 intensity score and the average microvessel density (Fig. 3d).
Ovarian cancer cell lines and primary ovarian tumors express hCAP-18/LL-37
Human ovarian tissue, including malignant and borderline tumors as well as normal tissue, was analyzed by Western blot and ELISA for expression of hCAP-18/LL-37. The tissue examined included a thecoma (Fig. 4a, lane 1), a granulosa cell tumor (lane 2), a serous papillary adenocarcinoma (lane 3), 3 serous borderline tumors (lanes 4, 6, 9), nonmalignant, matched ovarian tissue taken from the a patient who also provided a serous borderline tumor sample (lane 5), a mucinous borderline tumor (lane 7), an endometrioid adenocarcinoma (lane 8), an ovarian tumor of unknown type (lane 10) and a carcinosarcoma (lane 11). A similar banding pattern to what has been previously reported was observed, including that for recombinant LL-37 peptide.15, 17, 32 Full-length hCAP-18/LL-37 pro-peptide was detected in all ovarian tissue tested. Malignant and borderline samples exhibited immunoreactive bands corresponding to one of the 2 recombinant peptide bands in every case. LL-37 was not detected in noncancerous ovarian tissue (lane 5). The granulosa cell tumor, a serous borderline tumor, endometrioid adenocarcinoma and carcinosarcoma demonstrated the highest expression of LL-37 peptide (lanes 2, 4, 8, 11). ELISA was performed to quantify and validate these results (Fig. 4, values under blot).
A panel of ovarian cancer cell lines representing different tumor types (see Materials and methods Section) was examined for hCAP-18/LL-37 expression by Western blot analysis. Peripheral blood mononuclear cell (PBMC) lysate was used as control. Full-length hCAP-18/LL-37 pro-peptide was found to varying degrees in all cancer cell lines examined, while the peptide LL-37 was notably absent (Fig. 4b). Hs832.Tc, a cell line derived from a benign ovarian cyst, and TOV-21G, a clear cell carcinoma, expressed the lowest hCAP-18/LL-37 protein levels. These data indicate that ovarian cancer cells constitutively express hCAP-18 under traditional cell culture conditions.
LL-37 promotes human ovarian cancer cell proliferation
Previous studies have shown that LL-37 stimulates cell growth.8, 11, 32 Thus, LL-37 was assessed for its ability to induce proliferation of 3 independent ovarian cancer cell lines. Cells were treated with varying doses of LL-37 or epidermal growth factor (EGF), as a positive control. After 48 hr, LL-37 effectively increased proliferation of all 3 ovarian cancer cell lines tested (Fig. 5). The difference between untreated and LL-37-treated groups was significant for HEY cells at 1 and 5 μg/mL, for OV-90 at 1 μg/mL, and SK-OV-3 at all 3 concentrations used (p < 0.05). Notably, LL-37 induced proliferation exclusively in the presence of serum (data not shown).
LL-37 stimulates the migration and invasion of ovarian cancer cells
Because LL-37 functions as a potent chemoattractant for subsets of immune cells and even some epithelia, the Boyden chamber migration assay was employed to determine if the peptide elicits a similar response from ovarian cancer cells.11, 23, 25, 26, 27 Serum-starved cells were added to the Boyden chamber containing chemoattractants in the presence of 0.5% bovine serum albumin (BSA) or 0.5% fetal bovine serum (FBS) to determine if serum factors are required for LL-37-stimulated migration. LL-37 induced the migration of both HEY and SK-OV-3 cells without the aid of serum, but the peptide's effects were further enhanced when combined with FBS (Fig. 6). LL-37 significantly increased SK-OV-3 cell migration at concentrations of 0.1 and 1 μg/mL (p < 0.05).
The ability of LL-37 to stimulate invasion of ovarian cancer cells was also tested using the FluoroBlok invasion assay. In this assay, Matrigel served as the extracellular matrix that ovarian cancer cells were challenged to invade. Serum-starved HEY and SK-OV-3 cells were seeded on Matrigel-coated transwell inserts that were loaded into medium containing varying doses of LL-37 or EGF in the presence of serum. As shown in Figure 7, LL-37 significantly stimulated SK-OV-3 cell invasion (p < 0.05); although, the optimal dosage of the peptide was higher (10 μg/mL) than that observed in the migration assay (0.1 or 1 μg/mL). HEY cells were relatively unresponsive to LL-37 and EGF in this assay most likely due to the highly invasive nature of these cells.
The results of the invasion assay suggested LL-37 might activate enzymes to degrade extracellular matrix. Therefore, it was of interest to measure matrix metalloproteinase (MMP) secretion and activity from LL-37-treated ovarian cancer cells. Conditioned medium was collected and analyzed by zymography from samples treated with 5 μg/mL LL-37 or 10 ng/mL EGF for 24 hr. LL-37 and EGF treatment of HEY cells increased the secretion of both the pro- and active forms of two gelatinases, MMP-2 and MMP-9 (Fig. 7c). The same increase in MMP-2 was observed after LL-37 and EGF treatment of SK-OV-3 cells, whereas MMP-9 remained nearly undetectable regardless of treatment. No expression or activity was noted in casein gels indicating that the ovarian cancer cells examined do not abundantly secrete stromelysins (data not shown).
Real-time PCR was used to quantify the increase in protease mRNA transcripts after LL-37 treatment of HEY and SK-OV-3 ovarian cancer cells. As shown in Figure 7d, LL-37 and EGF significantly stimulated the expression of MMP-2, MMP-9, MMP-14 (MT1-MMP) and urinary plasminogen activator (uPA) in SK-OV-3 cells (p < 0.05). Protease expression was not significantly increased by LL-37 or EGF treatment of HEY cells because of the high basal levels of protease mRNA in these cells.
A number of epidemiological studies have recently recognized the causal relationship between chronic inflammation and ovarian cancer development.2, 3, 30 However, the molecular mechanisms that govern this process are not well understood. In the present study, we have characterized the expression of hCAP-18/LL-37 in normal and cancerous ovarian tissue and investigated a potential role for this pro-inflammatory peptide in ovarian tumors. Here, we demonstrate a number of important and novel observations: the pro-inflammatory, antimicrobial molecule LL-37 and its pro-form hCAP-18 are significantly overexpressed in ovarian tumors when compared to normal ovarian tissue; the expression of hCAP-18/LL-37 correlates with immune cell infiltration and angiogenesis in epithelial-derived ovarian tumors; and the peptide LL-37 induces ovarian cancer cell growth, chemotaxis, invasion and MMP secretion. Taken together, these data suggest that LL-37 promotes ovarian tumor progression through direct stimulation of tumor cells, initiation of angiogenesis and recruitment of immune cells.
Our data demonstrate a dramatic change in hCAP-18/LL-37 expression between normal and cancerous ovarian tissue. Normal tissue exhibited localized staining mainly by leukocytes whereas hCAP-18/LL-37 distribution in cancerous tissue was more widespread across various cell types (Fig. 2). Overexpression of hCAP-18/LL-37 in ovarian cancers is logical considering its established role in angiogenesis and inflammation; however, the function of LL-37 in normal ovary remains unclear.6, 8, 13, 17 LL-37 is known to initiate tissue repair and thus could be involved in reepithelialization of the ovarian surface following ovulation, a recurrent inflammatory process.11, 27 Other cytokines and growth factors with similar roles have been shown to be important in normal ovarian functions.33 Therefore, we would speculate that leukocytes and granulosa cells secrete hCAP-18/LL-37—since both cell types express the protein (Fig. 2)—to aid in the regulation of normal processes such as follicle growth and differentiation, ovulation and corpus luteum formation.
Western blot analysis of ovarian tumor tissue lysates revealed a banding pattern consistent with other reports.17, 34 The antibody used in this assay recognizes the LL-37 peptide, the hCAP-18 pro-form and the prepro-form all of which were present in ovarian tumor tissue (Fig. 4). LL-37 was not detected in normal ovarian tissue. Recombinant LL-37 displayed two distinct immunoreactive bands indicating that the peptide may oligomerize and evidence exists to support this notion.35, 36 Because serine proteases are responsible for releasing the LL-37 peptide from its precursor hCAP-18 extracellularly, it would be interesting to determine which proteases cleave hCAP-18/LL-37 in the ovarian tumor microenvironment. Recently, 2 members of the kallikrein family, KLK5 and KLK7, were found to control processing of hCAP-18 and LL-37 in skin.37 These same two proteases are also highly expressed in epithelial ovarian tumors when compared to normal ovarian tissue and they contribute to ovarian cancer progression.38, 39, 40 Thus, KLK5 and KLK7 are potential modulators of the pro-inflammatory activity of LL-37 in ovarian tumors.
Statistical analysis of hCAP-18/LL-37 expression with CD45 or CD31 expression in epithelial-derived ovarian tumors led to 2 interesting observations (Fig. 3). A significant positive correlation was found between hCAP-18/LL-37 and CD45 expression suggesting that tumor-derived hCAP-18/LL-37 functions to recruit immune cells to ovarian tumors. This observation is supported by previous studies demonstrating the peptide's chemotactic actions on neutrophils, monocytes, and CD4+ T cells, but not myeloid-derived dendritic cells or CD8+ T cells.23, 26, 31 Secondly, hCAP-18/LL-37 expression was significantly correlated in epithelial-derived tumors displaying increased microvessel density. LL-37 has been shown to function as a pro-angiogenic factor; thus, the peptide most likely stimulates angiogenesis in ovarian tumors.6, 17
In addition to its proposed actions in leukocyte recruitment and angiogenesis, we report here that LL-37 may also affect ovarian cancer cells directly. LL-37 treatment of ovarian cancer cell lines augmented cell growth, but only in the presence of serum (Fig. 5). Previously published reports have observed the same phenomenon, which raises the possibility that LL-37 needs accessory or carrier proteins to execute its effects.8, 11, 32, 41 In support of this notion, LL-37 has been shown to bind a number of molecules including glycosaminoglycans and mucins.42, 43, 44, 45 However, the peptide's ability to stimulate chemotaxis was not dependent on serum suggesting that LL-37's induction of proliferation and chemotaxis occurs through different mechanisms (Fig. 6). EGF receptor transactivation has been implicated in these processes in other cell types.11, 19, 27 Thus, it will be of interest to determine if LL-37 is acting through similar EGF receptor-dependent mechanisms to activate ovarian cancer cells.
Our in vitro studies demonstrating LL-37's stimulation of chemotaxis, invasion, and MMP secretion indicate that the peptide increases the metastatic potential of ovarian cancer cells. Release and activation of MMPs by tumor-derived LL-37 may serve to facilitate extracellular matrix degradation and metastasis of tumor cells. LL-37 has been shown to activate proteases that cleave membrane-anchored EGF receptor-ligands in other cell types.11, 19, 27 Therefore, it is not surprising that the peptide upregulated MMP-2 and MMP-9 in ovarian cancer cells (Fig. 7). Because these 2 MMPs are established promoters of tumor progression, our data further emphasize a pro-tumorigenic role for LL-37 in ovarian cancer.46
In conclusion, the results of this study present novel data demonstrating the upregulation of hCAP-18/LL-37 in ovarian tumors and begin to describe the peptide's importance in ovarian cancer biology. Although it will be essential to establish hCAP-18/LL-37's role in ovarian tumor development and progression using in vivo models, our findings provide further insight into the significance of pro-inflammatory molecules in ovarian cancer progression.
We thank Dr. Frank Marini for generously supplying OVCAR-3 and HEY ovarian cancer cell lines, Dr. Deborah Sullivan for technical assistance, and Dr. Lisa Morici for critical review of the manuscript.