OVCAR8 cells and their derivative lines were cultured in Dulbecco-Modified Eagle’s Medium with 10% fetal bovine serum and 1% penicillin/streptomycin, in a humidified incubator at 37°C/5% CO₂. OVCAR8 cells stably expressing null-vector control plasmid (OVCAR8-NV) and HE4 overexpression plasmid (OVCAR8-C5) were established as previously described, with OVCAR8-C5 cells secreting HE4 levels > 800 pM (12). OVCAR8 wild-type (WT) cells were treated with 20 nM human recombinant HE4 (rHE4; MyBioSource, MBS355616) or 50% conditioned media from OVCAR8-C5 cells for various lengths of time, as indicated.

Subconfluent OVCAR8-WT cells were treated with rHE4 in triplicate for 6 h, and RNA was isolated by Trizol extraction/LiCl precipitation. Subconfluent OVCAR8-NV and OVCAR8-C5 cells were collected in triplicate and RNA isolated in the same manner. Purity was determined by NanoDrop 2000 (Thermo Scientific), and RNA integrity was measured by Bioanalyzer (Agilent 2100) before microarray analysis (Affymetrix HuGene-1_0-st-v1) of 150 ng starting material at the Brown University Genomics Core Facility. Raw intensity values were converted to CELS files, and Transcriptome Analysis Console (TAC) Software was used to generate fold changes and ANOVA p-values. ANOVA p-values < 0.05 were considered significant. The top 15 genes in either direction were determined for OVCAR8-WT cells (untreated/rHE4-treated) and OVCAR8-NV/OVCAR8-C5 cells. Genes that were changed in both OVCAR8-WT (untreated/rHE4-treated) and OVCAR8-NV/OVCAR8-C5 comparisons with fold change greater than 1.5 in either direction were determined.

The DAVID v6.7 (22, 23) was used to identify the top four enriched annotation terms among genes differentially expressed (1.5-fold in either direction, p < 0.05) between OVCAR8-WT (untreated/rHE4-treated) and OVCAR8-NV/OVCAR8-C5, as well as among overlapping genes in these two comparisons. Default DAVID parameters were employed as follows:

Kappa Similarity: Similarity Term Overlap—3; Similarity Threshold—0.5

RNA was isolated by Trizol extraction/LiCl precipitation. Total RNA (500 ng) was reverse transcribed into cDNA using an iScript cDNA Synthesis Kit (Bio-Rad, 1708890) according to the manufacturer’s protocol. For microarray validation, the same RNA samples were used. 1 µL cDNA reaction, 2 µL each of 5 µM custom forward and reverse primers (Invitrogen, Sino Biological Inc.) or 1 µM forward and reverse-validated primers (http://realtimeprimers.com), 10 µL SYBR Green (Applied Biosciences [ABI], 4367659), and 5 µL RNAse-free water were added to each well of a 96-well plate for qRT-PCR analysis. Plates were run on an ABI 7500 Fast Real-Time PCR System, and data were analyzed using the ΔΔCt method. Relative expression levels were normalized to 18S rRNA to correct for equivalent total RNA levels. Validated LAMC2 and LAMB3 primers were purchased from http://realtimeprimers.com. Validated SERPINB2 primers were purchased from Sino Biological Inc. (HP100614). Custom primer sequences (Invitrogen) are as follows:

GREM1—F-GGGAGCCCTGCATGTGAC

Protein was extracted in Cell Lysis Buffer (Cell Signaling, 9803) with 1 mM PMSF, and concentrations were determined by DC Protein Assay (Bio-Rad Laboratories, 5000116). Equal amounts of protein boiled with Novex Sample Reducing Agent (Life Technologies, NP009) and NuPAGE LDS sample buffer (Thermo Fisher Scientific, NP0007) were loaded into a 4–12% gradient NuPAGE Novex Bis-Tris gel [Life Technologies, NP0321BOX (mini) and WG1402BX10 (midi)]. Protein was transferred by semi-dry transfer to methanol-activated 0.2 µm PVDF membranes (Bio-Rad, 162-0177) at 0.12–0.2 A for 1 h. Blocking was performed in 5% milk in phosphate-buffered saline with 0.05% Tween 20 (PBS-T) for 30 min at room temperature. Membranes were incubated in primary antibody in 5% milk in PBS-T overnight at 4°C and then in secondary antibody in 5% milk in PBS-T for 1 h at room temperature, with PBS-T washes in between. HRP-tagged secondary antibodies were detected by Amersham ECL Prime Western Blot Detection System (GE Healthcare, RPN2232). Blots were imaged directly in a Bio-Rad ChemiDoc MP Imaging System. GAPDH was used as a loading control. Original images can be seen in Figure S1 in Supplementary Material. Antibodies and dilutions used are as follows:

LAMC2 (Santa Cruz, sc-28330, 1:200)

Densitometry analysis of Western blots was performed using Image J. Blots were analyzed in eight-bit TIFF format with the “analyze gel” function. Band densities were normalized to GAPDH or the appropriate total protein for phosphoproteins. The lowest value was set to 1 for plotted graphs.

OVCAR8-WT cells were treated with 50% OVCAR8-C5 conditioned media for 48 h or left untreated. Protein was collected using lysis buffer provided in the Proteome Profiler Human Phospho-Kinase Array Kit (R&D Systems, ARY003B). The manufacturer’s instructions for the kit were followed, and membranes were developed in a Bio-Rad ChemiDoc MP Imaging System. Image J was used to perform background subtraction and determine spot density.

For one replicate of the invasion assays, a Cytoselect 24-Well Cell Invasion Kit (8 μm, Colorimetric, Cell Biolabs, CBA-110) was used according to the manufacturer’s instructions. After overnight starvation, OVCAR8 cells (1 × 10⁵/well) were plated in triplicate in serum-free media in cell culture inserts in the presence or absence of 20 nM rHE4. Media containing 10% FBS were inserted into the lower chamber. After 24 h, media were aspirated from the insert, and the top side of the insert was cleaned with a cotton swab. The insert was then crystal violet stained and washed, and the stained cells were extracted. Extraction solution (150 µL) from each sample was then transferred to a 96-well plate and OD measured at 550 nm. For the following two experimental replicates, 8 µm Transwell Permeable Supports Coated with Cultrex BME (Corning Inc., 3458) were used, with crystal violet staining and acetic acid extraction.

Cytoselect 48-Well Cell Adhesion Assay Kit (Fibronectin-Coated, Colorimetric; Cell Biolabs, CBA-050) was used to determine the effect of 20 nM rHE4 treatment on adhesion of OVCAR8-WT cells. Cells were plated in triplicate in serum-free media at 1 × 10⁵/well with or without 20 nM rHE4 in the assay plate for 2 h. Media were aspirated, and cells were stained, washed, and extracted. Extraction solution (150 µL) from each sample was transferred to a 96-well plate and OD read at 550 nm.

Haptotaxis assays were conducted using Transwell plates (6.5 µm thickness, 8 µm pores; Corning Inc., 3422). The lower surfaces of the Transwell membranes were coated by adding 500 µL of serum-free PRF-DMEM/F12 containing 2 µg/mL human fibronectin to the lower reservoir overnight. OVCAR8 cells in serum-free PRF-DMEM/F12 were seeded into the upper reservoirs of the Transwell inserts in the presence or absence of 20 nM HE4 and allowed to migrate overnight. Non-migrated cells were removed from the upper surface of the membrane using a Q-tip, and the cells attached to the lower surface were stained with 0.4% crystal violet in sodium borate buffer, pH 9.2 for 5 min, and then washed 2× in water. Crystal violet was eluted from the cells using acetic acid and measured spectrophotometrically at 550 nm.

A SensoLyte Rh110 Matriptase Activity Kit (AnaSpec, 72241) was used according to the manufacturer’s protocol. Various doses of rHE4 (10, 20, and 40 nM) were tested, and fluorescent readouts (490/520 nm) were performed after 30 min incubation at 37°C. Leupeptin was used as an inhibitor control. Average fluorescence from the substrate control wells was subtracted from the rHE4 test control wells, and the remainder was subtracted from the values obtained for the rHE4-treated wells.

To visualize focal adhesions, OVCAR8 cell plated on glass coverslips were starved in phenol red-free, serum-free media for 48 h. Starved cells were treated with 1 µg/mL rat fibronectin (Millipore, 341668) in the absence or presence of 20 nM HE4 for 2 h at 37°C. Cells were then washed with PBS and fixed in 4% paraformaldehyde for 10 m and permeabilized in 0.1% Triton X-100 for 5 m. Cells were blocked in 4% normal horse serum in PBS for 30 m and then incubated with phosphotyrosine clone 4G10 antibody (Millipore, 05-321) in 4% normal horse serum (1:500) for 1 h at room temperature. Coverslips were washed in PBS, and cell-associated antibodies were detected using DyLight 594 anti-mouse secondary (Vector Laboratories, DI-2594) diluted in 4% normal horse serum (1:1,000) for 45 min at room temperature. After staining, coverslips were washed and mounted on glass slides in Vectashield with DAPI (Vector Laboratories, H-1200). Imaging was performed on a Zeiss Axio Imager M1 using associated AxioVision software. Quantification was performed in Image J by measuring mean gray values of 4–6 fields per replicate after background subtraction (rolling ball radius = 50 pixels, sliding paraboloid). Results are the average of three independent experiments with 1–3 replicates each.

For laminin-332 staining, OVCAR8-WT cells were seeded onto glass coverslips and treated the next day with 20 nM rHE4 for 5 h. Cells were washed with PBS, fixed in 4% paraformaldehyde for 10 min, and permeabilized in 0.1% Triton X-100 for 5 m. Blocking was performed in 4% normal horse serum for 30 min, and the cells were incubated with anti-laminin-332 antibody (Abcam, ab14509) diluted in 4% normal horse serum (1:200) overnight at 4°C. The following day, the coverslips were washed in PBS and incubated in DyLight 488 anti-rabbit secondary (Vector Laboratories, DI-1488) diluted in 4% normal horse serum (1:1,000) for 45 min at room temperature. Coverslips were mounted using Vectashield with DAPI (Vector Laboratories, H-1200). Imaging and quantitative analysis were performed by an experienced technician at the RI Hospital Core Digital Imaging Facility. Results are the average of three independent experiments with three replicates each.

Where statistics are shown, n ≥ 3 independent experiments with biological replicates ≥3, and p-values were determined by unpaired one-tailed Student’s t-test. Transcriptome Analysis Console (TAC) software (Affymetrix) was used to generate fold changes and ANOVA p-values for microarray data.

To evaluate the effect of HE4 on malignant characteristics of OVCAR8 cells, we performed assays to evaluate invasion, adhesion, and migration. When invasion capacity was evaluated, OVCAR8-WT cells treated with rHE4 were found to have a 2.07-fold greater invasion capacity than untreated cells (p = 0.010788). Treatment with rHE4 also increased adhesion of cells to a fibronectin matrix by 1.29-fold (p = 0.002257). Interestingly, adhesion onto other substrates—collagen I and IV, laminin I, and fibrinogen—was not altered (data not shown), revealing the specificity of HE4’s effect on adhesion to fibronectin. We then tested the effect of rHE4 treatment of OVCAR8 cells on haptotaxis toward a fibronectin substrate. Treatment with rHE4 increased haptotaxis of the cells by 1.72-fold (p = 0.000378) (Figure 1). Collectively, our results demonstrate that HE4 has a direct effect on metastatic properties of ovarian cancer cells.

To investigate the effect of recombinant HE4 treatment on transcriptional regulation in OVCAR8 ovarian cancer cells, we treated subconfluent cells with 20 nM rHE4 for 6 h. Simultaneously, we collected RNA from OVCAR8 cells stably overexpressing HE4 (C5) or a null vector plasmid (NV). We have previously shown that stable overexpression of HE4 in OVCAR8-C5 cells produces HE4 levels in conditioned media of >800 pM (12). Serial dilution revealed an HE4 concentration of 504 pM in media diluted 1:10 and 51 pM in media diluted 1:100 (data not shown). Total RNA was submitted for microarray analysis comparing OVCAR8-NV to OVCAR8-C5 and OVCAR8-WT untreated cells to OVCAR8-WT rHE4-treated cells. The top 15 annotated, protein-coding genes significantly changed in either direction were identified, as shown in Table 1. The complete results are available through ArrayExpress under the accession number E-MTAB-6366.

To narrow down our genes of interest, we then identified all transcripts that were changed ≥1.5-fold in either direction by both HE4 overexpression and rHE4 treatment (Table 2). Six of these genes were regulated in opposite directions (DENN/MADD domain containing 2 A [DENND2A], family with sequence similarity 19, member 1A [FAM19A1], dehydrogenase/reductase member 3 [DHRS3], transgenlin [TAGLN], leucine rich repeat interacting protein 1 [LRRFIP1], and interleukin 6 [IL6]), suggesting that stable overexpression of HE4 has some different effects on transcription than short-term exposure of cells to HE4 protein. While we found a few transcripts that were downregulated by HE4, many upregulated genes were of particular interest because of their involvement with invasion and metastasis in diverse tumor types.

Several genes that were significantly upregulated code for extracellular matrix proteins, such as serpin peptidase inhibitor, member 2 (SERPINB2), gremlin 1 (GREM1), laminin-β3 (LAMB3), laminin-γ2 (LAMC2), fibroblast growth factor 5 (FGF5), tenascin C (TNC), adherens junctions associated protein 1 (AJAP1), and growth and differentiation factor 6 (GDF6). We validated the upregulation of SERPINB2, GREM1, LAMB3, LAMC2, and TNC by qRT-PCR (Figures 2A,B). SERPINB2 was upregulated by 2.13-fold (p = 0.024991) and 3.31-fold (p = 0.015549) by HE4 stable overexpression and rHE4 treatment, respectively; LAMB3 was upregulated by 4.04-fold (p = 0.0004) and 7.02-fold (p = 0.00022), respectively; LAMC2 was upregulated by 4.45-fold (p = 0.002671) and 5.46-fold (p = 0.00079), respectively; GREM1 was upregulated by 3.19-fold (p = 0.015549) and 4.94-fold (p = 0.0000241), respectively; and TNC was upregulated by 3.86-fold (p = 0.028017) and 6.06-fold (p = 0.002942), respectively.

Gene ontology analysis of the differentially expressed transcripts revealed enrichment of terms related to the extracellular matrix, cell migration, adhesion, and growth (Table 3). We furthermore noted that treatment of OVCAR8-WT cells with rHE4 promoted enrichment of gene terms related to phosphorylation/protein kinase activity, suggesting that addition of exogenous HE4 may have an effect on protein kinase cascades.

We chose to further investigate HE4’s regulation of LAMC2 and LAMB3, since these two genes code for chains of laminin-332, a secreted heterotrimer that has been well-described to promote aggression and metastatic properties in diverse cancers (24–31). LAMC2 and LAMB3 protein levels were constitutively elevated in OVCAR8-C5 cells compared to OVCAR8-NV (Figures 3A–C), while their levels peaked between 4 and 24 h after treatment with both rHE4 and conditioned media from OVCAR8-C5 cells (Figures 3D–G). Interestingly, C5 media (which has an approximate HE4 concentration of 5 nM) elicited a stronger response than 20 nM rHE4, suggesting that naturally secreted HE4 may be more bioactive than recombinant protein. Collectively, these results reveal a time-dependent effect of exogenous HE4 on LAMC2 and LAMB3 and indicate that stable overexpression of HE4 leads to constitutively high levels of these two proteins.

Next, we performed immunofluorescence analysis of the complete laminin-332 heterotrimer to determine if overexpression of its subunits promotes increased levels of the complete heterotrimer. Laminin-332 staining (mean gray value) was increased by 1.36-fold with C5 media treatment (p = 0.069025) and by 1.48-fold with rHE4 treatment (p = 0.002926), indicating that elevated levels of LAMC2 and LAMB3 proteins resulted in increased secretion of the laminin-332 heterotrimer (Figures 3H,I).

Since HE4 has been shown to inhibit the activity of multiple proteases (32, 33), we hypothesized that it might also inhibit enzymatic activity of matriptase, a serine protease that is known to proteolytically cleave laminin-332 in its β3 chain. Surprisingly, we observed the opposite to be true. In vitro matriptase activity was enhanced by rHE4 in a dose-dependent manner, while the inhibitor control (leupeptin) almost entirely obliterated matriptase activity (Figure 4). When the results of 3–5 separate experiments per dose were averaged, we saw an average 1.26-fold (±0.534075) increase in activity by 10 nM rHE4 (p = 0.04139), 1.32-fold (±0.489149) by 20 nM rHE4 (p = 0.000798), and 1.63-fold (±0.61337) by 40 nM rHE4 (p = 0.026596) (data not shown). Although these results are in opposition to the presumed role of rHE4 as a protease inhibitor, they indicate that HE4 not only upregulates laminin-332 levels but may also contribute to regulation of its proteolytic processing, which is known to promote migration in prostate cancer cells (34).

Next, we pursued the hypothesis that HE4 promotes activation of diverse kinase signaling pathways. The activation of MAPK signaling by HE4 has been well documented by us and others (8, 10, 13, 14), but we suspected that activation of other signaling pathways is also stimulated by HE4. We treated OVCAR8-WT cells with conditioned media from OVCAR8-C5 cells for 48 h and analyzed phosphorylation of an array of kinases and their target proteins using a Human Phospho-Kinase Array (R&D Systems; Figure 5A). Upregulation of several phosphoproteins was observed, including phospho-ERK (2.34-fold), as we have previously reported (8). Of note, we also found β-catenin and phospho-FAK (Y397) upregulated by 3.13- and 2.76-fold, respectively. β-Catenin is an intracellular signal transducer of the Wnt-signaling pathway, which has roles in cell–cell adhesion (35), and FAK is a Src-family tyrosine kinase that promotes cell adhesion to the extracellular matrix (36). Furthermore, multiple other members of the Src-family of tyrosine kinases, including phospho-Fgr (Y412), phospho-Fyn (Y420), phospho-Hck (Y411), phospho-Src (Y419), and phospho-Yes (Y426), were all upregulated by around 1.5-fold or more in response to exogenous HE4 exposure.

To confirm the effects of HE4-mediated activation of FAK signaling, we stimulated starved OVCAR8-WT cells with rat fibronectin in the presence or absence of rHE4 for 2 h and examined focal adhesions by phosphotyrosine staining. Cells that were treated with rHE4 had 1.61-fold more focal adhesions than untreated cells (p = 0.0000593), indicating that HE4 phosphorylation of FAK at Y397 promotes subsequent formation of focal adhesions (Figures 5B,C).