Cell lines and Growth Conditions

The Wistar Melanoma (WM) cell lines were kindly provided by Dr. Meenhard Herlyn (Wistar Institute, Philadelphia, PA) and have been described in detail elsewhere [13].

The MeWo cell line was derived from a lymph node metastasis [14]. The cell lines FEMX-I and LOX were established from metastatic lymph node biopsies obtained from melanoma patients treated at the Rikshospitalet-Radiumhospitalet Medical Center [15]. The cells were routinely cultured in RPMI 1640 medium (BioWhittaker Europe, Verviers, Belgium) supplemented with 5% fetal calf serum (FCS) (Biochrom, KG, Berlin, Germany). Phorbol-12-myristate-13-acetate (PMA) was from Sigma-Aldrich (St. Louis, MO), whereas the MEK1 inhibitor, PD98059, was from Cell Signaling Technology (Beverly, MA). Multi-cellular aggregates (spheroids) were prepared as previously described [16]. Briefly, 24-well plates were coated with 1% Seaplaque agarose (BioWhittaker Molecular Application, Rockland, ME) and tumor cells (2 × 10⁵ cells in 1 ml complete medium) were plated on top of the solidified agarose. For thymidine incorporation assay, 5000 cells per well were plated in 96-well polyhema (Sigma-Aldrich)-coated U-bottom plates. For treatment of spheroid cultures, PMA was added when plating in suspension, whereas the inhibitors in combination experiments were added 45 min prior to plating as spheroids.

Gene expression analysis

WM35 cells were grown as spheroids for 24 hrs in the presence of PMA and PD98059, alone and in combination. Total RNA was extracted using the TRIZOL reagent (Invitrogen, Carlsbad, CA). Gene expression profiling was performed using Affymetrix U133 Plus 2.0 arrays (Affymetrix, Santa Clara, CA). For microarray hybridization, the protocol described in the Affymetrix GeneChip eukaryotic one-cycle target preparation protocol, using 5 μg of total RNA, was followed. Analysis of the data was performed by Genolyze Ltd. (Turku, Finland) using statistical software R version 2.3.0. and package collection Bioconductor version 1.8. Statistical significance was assessed using p-value from two-tailed two sample t-test. P-values are replaced with q-values to control the False Discovery Rate.

Quantitative real time RT-PCR analysis

The high capacity cDNA reverse transcription kit (Applied Biosystems, Foster city, CA) was used to reverse-transcribe total RNA (0.8 μg) in a 20 μl reaction mixture using random primers. The real-time PCR analyses were performed using TaqMan Fast Universal PCR Master Mix (2×) and TaqMan Gene Expression Assay (HS00361426-ml FABP7, HS99999908-ml GUS, Applied Biosystems). A total of 0.5 μl cDNA was used in 25 μl PCR mixtures with 900 nM of each primer and 250 nM TaqMan probe. The reactions were carried out in a 7900 HT Fast Real Time PCR system (Applied Biosystems) with the following program: 95°C for 20 sec. followed by 40 cycles of 95°C for 1 sec., 60°C for 20 sec. Each sample was run in triplicate. The FABP7 relative mRNA expression level was normalized with respect to the beta-glucuronidase (GUS) gene, which had stable transcript levels under these experimental conditions. The mean from three independent experiments was calculated.

Immunoblotting

Cells were lysed in ice-cold NP-40 lysis buffer (1% NP-40, 10% glycerol, 20 mM Tris-HCl, pH 7.5, 137 mM NaCl, 100 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride (PMSF), 0.02 mg/ml each of aprotinin, leupeptin, and pepstatin, and 10 μl/ml phosphatase inhibitor cocktail I and II (Sigma-Aldrich)). Protein quantitation was done by Bradford analysis and 25 μg protein/lane was resolved by SDS polyacrylamide gel electrophoresis. Transfer and hybridization were as described in [17]. To ensure even loading, filters were stained with naphthol-blue black (Sigma-Aldrich) and re-stained with α-tubulin. The antibodies against FABP7 and α-tubulin were from R&D Systems (Minneapolis, MN) and Calbiochem (San Diego, CA), respectively. HRP-conjugated anti-mouse IgG secondary antibody was from Promega (Madison, WI) and HRP-conjugated anti-goat secondary antibody was from DAKO A/S (Glostrup, Denmark).

Small interfering RNA transfection

Fifty thousand cells per well were seeded in 24-well plates for 24 hrs prior to transfection with 50 nM siRNA targeting FABP7 (OligioID: HSS103516; Catalog# 1299003) or negative control siRNA duplexes (Catalog#12935-300) using Lipofectamine™ RNAiMAX transfection reagent (all reagents and siRNA were from Invitrogen). Cells were detached 48 hours after transfection and plated into agarose-coated 24-well plates as spheroids for an additional 72 hrs for assessment of apoptosis, seeded into 96-well polyhema-coated U-bottom plates for the proliferation assay and plated in BioCoat Matrigel invasion chambers.

Proliferation assay

Five thousand cells per well were seeded in 96-well polyhema-coated U-bottom plates for spheroids and in 96-well flat-bottom plates for monolayer cells and cultured for 72 hrs, the last 24 hrs with the addition of 3.7 × 10⁴ Bq [³H]Thymidine (ARC, St.Louis, MO) Thereafter, the cells were harvested using a Filtermate Harvester (Packard Instrument Co. Meriden, CT). [³H]Thymidine incorporation was assessed in a Packard Microplate Scintillation Counter. Proliferation assays were measured in triplicate. The experiment was repeated at least three times.

Flow cytometric analysis of apoptosis

The adherent cells were harvested by Trypsin and together with detached cells fixated in 100% cold methanol. Fixed cells were washed with PBS, incubated for 30 min at 37°C in 50 μl terminal transferase (TdT) solution containing 5 units TdT (Roche, Basel, Switzerland), 10 μl 5× reaction buffer (supplied with TdT), 1.5 mM CoCl₂, 0.5 nmol labeled biotin-16-dUTP, 0.1 mM dithiothreitol and distilled water. The cells were subsequently washed once in PBS containing 0.1% Triton X-100 and incubated in 50 μl 1:50 streptavidin-FITC (Amersham, Buckinghamshire, UK) in PBS (0.1% Triton X-100) and 3% skimmed dry milk for 45 min at room temperature. After washing in PBS (0.1% Triton X-100) the pellet was resuspended in PBS (0.1% Triton X-100) containing 2 μg/ml Hoechst 33258 to a final concentration of 1 × 10⁶ cells/ml and incubated for 30 min at 4°C. Data acquisition and analysis were performed on Becton Dickinson LARII (Becton Dickinson immunocytometry systems, San Jose, CA) using Multifit software (FACSDiVa House inc., Tonsham, ME).

Matrigel invasion assay

WM35 and WM239 cells were plated in BioCoat Matrigel invasion chambers (BD Biosciences, San Jose, CA) at a cell density of 3 × 10⁴ per chamber in RPMI 1640 supplemented with 5% fetal bovine serum (inner chamber) 48 hrs post-transfection. Self-supplied fibroblast conditioned medium was used as chemoattractant in the outer chamber. The conditioned medium was obtained from fibroblasts isolated as described by Costea et al[18] cultivated in DMEM supplemented with 10% fetal bovine serum. The medium was collected when the cells were 70% confluent. After 48 hrs incubation at 37°C and 5% CO₂, non-invading cells remaining on the top surface of the chamber were removed by scrubbing with a cotton-tipped swab, and the invading cells that had adhered to the bottom surface of the chamber membranes were fixed, stained with hematoxylin and counted.

Clinical melanoma specimens

Formalin-fixed, paraffin-embedded tissue from 149 primary and 68 metastatic melanomas, as well as 11 benign nevi, was examined for expression of FABP7 protein. Of the primary tumors, 93 were classified as superficial spreading (SSM) and 56 as nodular melanomas (NM). Clinical follow-up was available for all patients. The study was approved by the Regional Committee for Medical Research Ethics in Norway.

Immunohistochemical analysis

Sections of formalin-fixed, paraffin-embedded tissue were immunostained using the two-step EnVision system (DAKO EnVision™, DAKO A/S). Deparafinized sections were microwaved in low pH buffer (pH 6.0) (DAKO) at 750 W for 5 minutes and then at 500 W for 15 minutes to unmask the epitopes. After treatment with 1% hydrogen peroxide for 5 minutes to block endogenous peroxidase, the sections were incubated with polyclonal rabbit anti-human FABP7 antibody (R&D Systems) for 30 minutes at room temperature followed by 30 minutes incubation with mouse anti-goat antibody (Santa Cruz Biotechnology, Santa Cruz, CA). The sections were then incubated with HRP-labeled secondary antibody for 30 minutes followed by 5 minutes incubation at RT with DAB substrate (DAKO A/S). All series included positive controls. Four semiquantitative classes were used to describe the number of stained cells: negative, ≤ 5%, 6–50% and >50%. Both nuclear and cytoplasmic staining was scored. Staining was evaluated by a surgical pathologist (BD). A subset of the cases (n = 50) was additionally scored by another author (AS).

Statistical analysis

Statistical analysis was performed using the SPSS program version 13.0 (Chicago, IL). The differences between FABP7 expression in benign nevi, primary melanomas and metastases were analyzed using the Chi-square test. The relationship between FABP7 expression and mean tumor thickness was evaluated nonparametrically using the Mann-Whitney two sample test. The association between expression of FABP7 and cell cycle markers was performed using the Fischer's exact test. Kaplan-Meyer estimates and the log-rank test were used for survival analysis. P < 0.05 was considered statistically significant.

Identification of molecules involved in survival of melanoma cells as multicellular aggregates in suspension using gene expression profiling

We previously showed that PMA treatment protects melanoma cells from suspension-mediated apoptosis while the MEK1 inhibitor PD98059 has the opposite effect [2]. In order to identify new factors involved in anchorage-independent growth of melanoma cells, we compared mRNA expression profiles from the melanoma cell line WM35, cultured in monolayer or as untreated spheroids, as well as following treatment of the spheroids with PMA and/or PD98059 for 24 hours.

The FABP7 gene was among the genes showing the highest differential expression. While no notable difference was observed between monolayer cells and spheroids, treatment with PMA or PD98059, as well as with PD98059 and PMA in combination, led to FABP7 mRNA down-regulation in treated spheroids compared to the spheroid control (Figure 1a). The microarray results were validated using real time RT-PCR (Figure 1b).

FABP7 is expressed in melanoma cell lines and regulated through PKC and the MAPK/ERK1/2 signaling pathway

The protein level of FABP7 in monolayer culture, untreated spheroids and spheroids treated with PMA and/or PD98059 for 24 hrs was analyzed using western blot. As shown in Figure 1c, no change in FABP7 protein level was observed between monolayer cells and untreated spheroids while in spheroids treated with PMA and/or PD98059, the protein level was reduced compared to controls. This was in accordance with the reduction of FABP7 mRNA levels. A similar reduction in FABP7 protein level was obtained in monolayer cultures treated with PMA and/or PD98059 (data not shown).

In order to reveal if FABP7 expression levels differ during the cultivation of the WM35 cells following PMA and/or PD98059 treatment, we performed a time course study. The monolayer cells were treated with PMA or PD98059 from 0,5 hrs to 72 hrs. As shown in Figure 2. we observed down-regulation of FABP7 protein after 12 hrs in both PMA and PD98059 treated cells although the effect of PMA was more pronounced over time. The down-regulation was sustained for up to 72 hrs for both treatments. These results were supported by real time RT-PCR (data not shown).

To examine if FABP7 is frequently expressed in melanoma cell lines we analyzed the level of FABP7 mRNA and protein in two primary (WM1341 and WM902B) and seven metastatic cell lines (WM239, WM45.1, WM983, WM9, LOX, MeWo and FEMX-I) in addition to WM35. As shown in Figure 3a and 3b, variable levels of FABP7 mRNA and protein were detected in 9 out of 10 cell lines. With the exception of WM45.1, good concordance between mRNA and protein levels was observed in all the cell lines. No clear differences were observed between FABP7 expression levels in cell lines originating from primary tumor vs. metastasis.

FABP7 is involved in proliferation and invasion of melanoma cells

In order to further investigate the function of FABP7 we chose to transiently down-regulate FABP7 using specific siRNA in the WM35 and WM239 cell lines, which we found to have high FABP7 expression. The effect of down-regulation on proliferation, invasion and apoptosis was examined. Monolayer cells were incubated for 48 hrs with FABP7 siRNA or a control siRNA and analyzed for transfection efficiency by western blot (Figure 4a). As demonstrated in figure 4b and 4c, FABP7 down-regulation reduced proliferation by 29% in WM35 and 84% in WM239 cells as compared to scrambled siRNA control transfected cells. Similar results were obtained when the cells were grown in suspension (data not shown).

The degree of apoptosis was assessed using TdT-mediated dUTP nick end labeling (TUNEL) staining and flow cytometry. Analysis of both monolayer and spheroid cultures showed that down-regulation of FABP7 did not affect the percentage of apoptotic cells (data not shown). Together these results suggest that FABP7 is most likely involved in proliferation and not apoptosis in melanoma cells.

We investigated the effect of FABP7 down-regulation on invasion using the Matrigel assay. The number of invading cells was reduced by 55% and 40% in WM35 and WM239 cell respectively after transfection with FABP7 siRNA compared with scrambled siRNA control-transfected cells (Figure 4d), suggesting that FABP7 contributes to the invasiveness of melanoma cells.

FABP7 is expressed in melanomas and associated with tumor thickness

In order to examine the clinical relevance of FABP7, paraffin-embedded tissue from a panel of benign nevi and primary and metastatic melanomas was analyzed for expression of FABP7 protein using immunohistochemistry. Heterogeneous cytoplasmic and/or nuclear expression of FABP7 was observed in 91% of the nevi, 71% of the primary tumors and 70% of the metastases. The results are summarized in Table 1 and 2 and illustrated in Figure 5. Statistical analysis demonstrated a significant higher cytoplasmic FABP7 expression in nevi compared to primary and metastatic melanomas (P = 0.023), with comparable nuclear expression. A two-tier analysis of primary and metastatic melanomas showed comparable expression for both cytoplasmic and nuclear expression (P > 0.05). Good concordance (>80%) was achieved between the two observers. Discrepant cases were resolved through a consensus session.

Tumor Type	No. of tumors	Total no. of positive	Cytoplasm	Nucleus	Cytoplasm/nucleus
Benign nevi	11	10 (91)	2 (18)	0 (0)	8 (73)
Primary melanomas	149	106 (71)	36 (24)	0 (0)	70 (47)
SSM	93	61 (66)	19 (21)	0 (0)	42 (45)
NM	56	45 (80)	17 (30)	0 (0)	28 (50)
Metastases	68	48 (70)	20 (29)	1 (2)	27 (39)

Tumor type	No. of tumors	Cytoplasm				Nucleus
		-	≤ 5%	6–50%	> 50%	-	≤ 5%	6–50%	> 50%
Benign nevi	11	1 (9)	0 (-)	1 (9)	9 (82)	3 (28)	4 (36)	4 (36)	0 (-)
Primary melanomas	149	43 (28)	14 (9)	42 (28)	50 (34)	79 (53)	51 (34)	16 (11)	3 (2)
SSM	93	32 (34)	10 (11)	25 (27)	26 (28)	51 (55)	31 (33)	9 (10)	2 (2)
NM	56	11 (20)	4 (7)	17 (30)	24 (43)	28 (50)	20 (36)	7 (12)	1 (2)
Metastases	68	21 (31)	8 (11)	21 (31)	18 (27)	40 (59)	20 (29)	7 (11)	1 (1)

Since 62% (92/149) of the primary tumors expressed cytoplasmic FABP7 in more than 5% of the cells, this cutoff was used to distinguish between high and low protein levels. Applying the same cutoff when evaluating nuclear staining, we observed that only 13% (19/149) of the tumors had high protein expression levels. Higher cytoplasmic FABP7 was significantly associated with increased thickness of SSM (P = 0.021). In addition, in this group of patients, a trend towards increased relapse-free survival (P = 0.069) for patients whose tumors expressed less FABP7 was observed. No such correlation was observed in NM (Table 3 and Figure 6). We did not observe any significant correlation between FABP7 staining and overall survival for patients diagnosed with either SSM or NM (data not shown). Nuclear staining had no association with disease outcome (data not shown).

Tumor type	Expression	No of patients	Average depth of growth (mm)	P
SSM	Negative*	41	1.59
	Positive	48	2.17	0.021
NM	Negative*	15	4.32
	Positive	41	4.78	0.697

* Considered negative if < 5% tumor cells showed positive immunoreactivety of FABP7

Relationship between FABP7 expression and markers of proliferation

Since our panel of primary and metastatic melanomas has previously been analyzed for expression of cell cycle progression markers [19–21] and activation status of MAPK/ERK1/2 [22], it was of interest to examine the relationship between FABP7 expression and the levels of these factors. The results showed no correlation between cytoplasmic FABP7 expression and the expression of the cyclins A, D1 or D3 or the cdk inhibitors p21^CIP1/WAF1and p27^Kip1 in either SSM or NM. However, a trend for an association between cytoplasmic FABP7 and Ki-67 (P= 0.07) in SSM was observed, which is in support of our in vitro results, suggesting FABP7 involvement in proliferation. Furthermore, expression of activated MAPK/ERK1/2 did not correlate with FABP7 expression. Interestingly, however, we observed that if the cutoff level was changed (only total lack of FABP7 was regarded as low expression), MAPK/ERK1/2 expression did positively correlate with cytoplasmic FABP7 expression in SSM (P= 0.047). This was not the case for NM, regardless of cutoff. Nuclear expression of FABP7 did not correlate with any of the examined markers (data not shown).