Aberrant Fatty Acid-Binding Protein-7 Gene Expression in Cutaneous Malignant Melanoma

Abstract

Fatty acid-binding protein-7 (FABP7) has been shown to be expressed in cutaneous melanoma; however, its role in tumor progression is unclear. Expression of FABP7 was assessed during melanoma progression through assessment of various clinicopathology stages of primary tumor progression and metastasis. FABP7 mRNA was highly expressed in 60 of 87 (69%) primary melanomas, compared with significant (P<0.0001) reduction in 13 of 68 (19%) metastatic melanomas. Analysis of 37 paired primary and metastatic melanomas by immunohistochemistry with anti-FABP7 Ab showed 73 and 27% positivity, respectively (P<0.001). FABP7 detection of metastatic tissues was inversely correlated with relapse-free (P<0.0001) and overall (P<0.0001) survival. To examine FABP7 expression loss in advanced melanomas, loss of heterozygosity (LOH) was assessed using microsatellite markers encompassing the FABP7 gene. LOH was identified in 10 of 20 (50%) metastatic melanomas at 6q22.31, compared with 0 of 14 primary melanomas (P = 0.0017). FABP7 as a surrogate biomarker for circulating tumor cells (CTCs) in the blood was assessed by quantitative real-time (qRT)–PCR from melanoma patients' blood (n = 134). Assessment of patients' blood showed that FABP7(+) CTC decreased with disease progression. FABP7 may function as a tumor progression gene and can be used as a potential diagnostic biomarker of early-stage melanoma systemic spreading in blood.

INTRODUCTION

Melanoma remains primarily a surgically treated disease that does not respond well to currently available systemic therapies, whereas early detection and diagnosis can significantly improve disease management and outcome (Morton et al., 2006). Better prognostic biomarkers of early-stage melanoma are needed to identify the potential of primary tumors to metastasize. Fatty acid-binding protein-7 (FABP7), also known as B-FABP (brain-type fatty acid-binding protein), and BLBP (brain lipid-binding protein) are involved in intracellular long-chain fatty acid transport (Sweetser et al., 1987; Feng et al., 1994; Kurtz et al., 1994; Shi et al., 1997). Originally identified in the mammalian brain, FABP7 has been linked to several cell-differentiation pathways that lead to the inhibition of cell proliferation and tissue differentiation (Anthony et al., 2005; Arai et al., 2005). FABP expression in tissues has been related to lipid-metabolizing capacity associated with fatty acids (Haunerland and Spener, 2004). There are several FABP family members, of which all are characterized by a β-barrel structure, membrane activity, and specific ligand-binding selectivity.

In cancer, FABP7 has been identified in glioblastoma (Liang et al., 2005), renal cell carcinoma (Domoto et al., 2007), and melanomas (de Wit et al., 2005; Goto et al., 2006). It has been shown that expression of FABP7 is correlated with survival in patients with glioblastoma (Liang et al., 2005), whereas FABP7 has also been shown to inhibit tumor growth of a breast cancer cell when overexpressed (Shi et al., 1997; Haunerland and Spener, 2004). However the role of FABP7 in tumor progression still remains unclear. In a recent study, FABP7 was identified as one of a panel of genes in the oligonucleotide arrays that appeared to be differentially expressed between nevus and metastatic melanoma in a small sample size study (de Wit et al., 2005).

We recently showed, by mRNA and western blot analysis, frequent expression of FABP7 in human metastatic melanoma cell lines and tissues as compared with various normal tissues. In addition, FABP7 was shown to be involved in proliferation and invasion of melanoma cells in vitro (Goto et al., 2006). The primary objective of this study was to determine FABP7 expression during melanoma progression, by assessing the expression of FABP7 mRNA in a large sample size of primary and metastatic melanomas of different American Joint Committee on Cancer (AJCC) stages. The secondary objective was to determine the association of FABP7 expression with survival of melanoma patients. Our group has previously shown the prognostic utility of tumor mRNA biomarkers in detecting and monitoring circulating tumor cells (CTCs) in the blood of patients with melanoma (Koyanagi et al., 2005a). FABP7 utility as a potential surrogate biomarker of CTCs in patients' blood comparing early and advanced stage melanoma was assessed.

RESULTS

FABP7 expression in tumors

FABP7 mRNA expression was assessed by an optimized quantitative real-time RT–PCR (qRT) assay in 155 cutaneous melanomas and 13 normal skin tissue. In the assessment of 155 melanoma specimens, the absolute mRNA copies of FABP7 ranged from 0 to 9.01 × 10². Receiver operating characteristic curve analysis was conducted to evaluate the diagnostic accuracy of FABP7 for the binary outcome (87 primary melanoma vs 13 normal skin tissue). The area under the receiver operating characteristic curve ± SEM was 0.824±0.032 (Figure 1).

Figure 1. Receiver operating characteristic of FABP7 in primary melanoma.

The AUC of FABP7 is 0.824 (P<0.0001).

To determine the pattern of FABP7 expression relative to early tumor progression, we examined the primary melanomas of specific AJCC stages of disease. FABP7 mRNA was detected in 20 of 30 (67%) AJCC stage I primary cutaneous melanomas, 17 of 25 (68%) AJCC stage II primary melanomas, and 23 of 32 (72%) AJCC stage III primary melanomas. These analyses showed that there was no significant differences in FABP7 expression in primary lesions regardless of the clinical stage. However, in advanced metastatic lesions, FABP7 was present only in 13 of 68 (19%) AJCC stage III and IV metastatic melanomas. The level of FABP7 mRNA expression was significantly higher in primary melanomas than in metastatic melanomas and in normal skin tissues (P<0.0001, P<0.0001, respectively, Figure 2). This indicated a decrease in FABP7 expression during tumor progression from primary to metastatic lesions, which suggested a significant downregulation of FABP7 in advanced stage metastatic melanomas.

Figure 2. Ratio of the FABP7/GAPDH mRNA copies in melanoma lines and tissues.

The cutoff point (dotted line) for FABP7 positivity was set at 0.0011 (the mean ratio of the FABP7/GAPDH mRNA copies plus 3 SD of normal human skin controls). Horizontal bars denote mean mRNA copies and SEM.

By immunohistochemistry (IHC) assessment, the expression of FABP7 protein using 37 paired tissues of primary and metastatic melanomas was analyzed. FABP7 protein was detected in 27 of 37 (73%) primary melanomas, whereas FABP7 was present only in 10 of 37 (27%) metastatic melanomas (P<0.0001; Supplementary Figure 1). Interestingly, the expression of FABP7 from 17 metastatic melanoma patients was lost, although paired primary melanomas from the same patients expressed FABP7 protein. This analysis further verified a decrease in FABP7 expression during tumor progression from primary to metastatic lesions.

Correlation of FABP7 expression with disease outcome

To assess whether FABP7 expression loss is a tumor progression prognostic biomarker, the correlation of FABP7 expression in metastatic tissues with relapse-free survival and overall survival in melanoma patients was assessed. Thirty-nine metastatic melanomas (19 AJCC stage III and 20 AJCC stage IV) were used for analyses of FABP7 expression correlation with disease relapse and overall survival. In the stage III group, there were 3 stage IIIa, 12 stage IIIb, and 4 stage IIIc. A median follow-up time was 31 months in all the patients. There were 4 of 19 (21%) stage III and 4 of 20 (20%) stage IV patients who expressed FABP7.

FABP7 detection was inversely correlated with relapse-free survival (P<0.0001; Figure 3a). There were 61 and 21% patients whose tumors did not express FABP7 and who were relapse free for 12 months and 60 months, respectively, and whose median relapse-free survival was 23.0 months (95% confidence interval (CI) 12.0–58.0). No patient with FABP7 positive survived relapse-free for 12 months, and the median survival months was 4.0 (95% CI 2.0–8.0). Univariate analysis also showed that pT stage (P = 0.014), M classification (P<0.0001), and AJCC stage (P = 0.013) were significantly correlated with relapse-free survival (Table 1A). Multivariate analysis showed that tumor FABP7 expression was a significant predictor for disease relapse (risk ratio, 6.47; 95% CI 2.47–16.96; P = 0.0001) as well as AJCC stage (risk ratio, 2.82; 95% CI 1.31–6.07; P = 0.0079, Table 1B).

Figure 3. FABP7 mRNA expression in metastatic tumors related to disease outcome.

Kaplan–Meier estimates of relapse-free survival (a) and overall survival (b) based on FABP7 mRNA detection in melanoma metastases.

Table 1A.

Univariate analysis of clinical factors for prediction of relapse-free survival and overall survival

		Relapse-free survival		Overall survival
Factors	Patients	Number	Log-rank	Number	Log-rank
Gender
Male	23	18	0.77	17	0.37
Female	16	12		9
Age (years)
≤50	18	15	0.84	13	0.92
>50	21	15		13
Primary site
Extremity	15	11	0.77	8	0.40
Head/neck	6	5		5
Trunk	12	9		8
Unknown	6	5		5
pT stage
pT1	8	8	0.014	6	0.20
pT2	8	2		2
pT3	7	6		6
pT4	10	9		7
pTO	6	5		5
Clark level
3	8	6	0.65	5	0.97
4	17	12		11
5	8	7		5
Unknown	6	5		5
N classification
N0	7	6	0.094	6	0.0096
N1	18	11		7
N2	10	9		9
N3	4	4		4
M Classification
M0	19	11	<0.00017	7	<0.0001
M1a	14	13		13
M1b	5	5		5
M1c	1	1		1
AJCC stage
Stage 3m	19	11	0.013	7	0.0008
Stage 4m	20	19		19
Expression of FABP7
Positive	8	8	<0.00018	8	<0.0001
Negative	31	22		18

Table 1B.

Multivariate analysis for disease outcome

	Relapse-free survival		Overall survival
Factor	P-value	Hazard ratio (95% CI)	P-value	Hazard ratio (95% CI)
AJCC stage	0.0079	2.82 (1.31, 6.07)	0.0006	4.782 (1.96, 11.66)
FABP7	0.0001	6.47 (2.47, 16.96)	<0.0001	14.52 (4.38, 48.09)

CI, confidence interval.

Moreover FABP7 expression was inversely correlated with overall survival (P<0.0001; Figure 3b). There were 87 and 43% patients whose tumors did not express FABP7 survived for 12 and 60 months, respectively, and whose median survival was 40.0 months (95% CI 24.0–103.0), whereas there were 38 and 0% patients with positive FABP7 who survived for 12 and 60 months, respectively, and whose median months of survival was 9.5 (95% CI 8.0–13.0). Univariate analysis also showed that N classification (P = 0.0096), M classification (P<0.0001), and AJCC stage (P = 0.0008) were significantly correlated with overall survival (Table 1A). Multivariate analysis showed that FABP7 detection (risk ratio, 14.52; 95% CI 4.38–40.09; P<0.0001) was a significant independent prognostic factor for overall survival as well as AJCC stage (risk ratio, 4.78; 95% CI 1.96–11.66; P = 0.0006).

FABP7 locus genomic aberration

To determine genomic deletion as the mechanism of the downregulation of FABP7 expression in metastatic melanoma, loss of heterozygosity (LOH) was assessed in the 6q21–23 chromosomal region that includes the FABP7 gene locus. Melanoma lines and Paraffin-embedded archival tissue (PEAT) tumor specimens were assessed for three defined microsatellite markers encompassing the FABP7 gene locus (Figure 4).

Figure 4. Location of FABP7 gene and the three microsatellite markers used for LOH assessment on chromosome 6.

FABP7 gene is located at 6q22.31, between D6S268 and D6S1702.

Nine cell lines established from metastatic melanomas were used for analysis (Table 2A). In four of nine cell lines (44%), LOH in the FABP7 gene region was identified in all three microsatellite markers; in one line, FABP7 LOH was identified in two of the three microsatellite markers. Interestingly, FABP7 mRNA expression detected by qRT was highly suppressed in four of five LOH cell lines, whereas FABP7 mRNA was expressed in three of four R cell lines (Figure 5) (Supplementary Figure 2). The inverse correlation between FABP7 LOH and mRNA expression suggested that deletion or genomic instability of the FABP7 locus during metastatic progression is a potential factor for loss or downregulation of the FABP7 gene. Cell line IHC analysis of FABP7 expression was concordant with LOH results except for one cell line.

Table 2A.

FABP7 LOH status in metastatic melanoma lines

	Microsatellite markers
Cell lines	D6S268	D6S1072	D6S262
ME-25	R	R	R
ME-26	LOH	LOH	LOH
ME-27	R	R	R
ME-28	R	R	R
ME-29	LOH	LOH	R
ME-30	R	R	R
ME-31	LOH	LOH	LOH
ME-32	LOH	LOH	LOH
ME-33	LOH	LOH	LOH

LOH, loss of heterozygosity; R, retention of alleles.

Figure 5. FABP7 mRNA expression, LOH status, IHC analysis of melanoma cell lines with polyclonal rabbit anti-FABP7 Ab.

Small bars represent SD. LOH, loss of heterozygosity; R, retention of alleles; +, IHC stained; −, IHC shows not stained; N/A, not available for analysis.

To confirm the LOH of the 6q21–23 observation in the cell lines, we assessed melanoma tumor tissues. We investigated LOH of the FABP7 region in 14 PEAT primary melanomas. As shown in Table 2B, FABP7 LOH frequency at individual microsatellite markers D6S268, D6S1702, and D6S262 was 14, 0, and 0%, respectively. By contrast, in 20 PEAT specimens from metastatic melanomas, the frequency of FABP7 LOH at D6S268, D6S1702, and D6S262 was 60, 50, and 50%, respectively (Table 2C); at least two of the three microsatellite markers were positive in 10 (50%) of these specimens. LOH was identified in 10 of 20 (50%) metastatic melanomas at 6q22.31, compared with 0 of 14 primary melanomas. The frequency of the 6q21–23 region LOH was significantly higher in metastatic melanomas than in primary melanomas (P=0.0017).

Table 2B.

LOH status in primary melanomas

		Microsatellite markers
Patient	Stage	D6S268	D6S1072	D6S262	LOH
1	I	R	R	R	N
2	I	R	R	R	N
3	I	R	R	R	N
4	I	R	R	R	N
5	I	R	R	R	N
6	I	R	R	R	N
7	II	R	R	R	N
8	II	LOH	R	R	Y
9	II	R	R	R	N
10	II	R	R	R	N
11	II	R	R	R	N
12	II	LOH	R	R	Y
13	III	R	R	R	N
14	III	R	R	R	N

LOH, loss of heterozygosity; R, retention of alleles.

LOH inpatient tumor: N, no; Y, yes.

Table 2C.

LOH status in metastatic melanomas

	Microsatellite markers
Patients	D6S268	D6S1072	D6S262	LOH
1	R	R	R	N
2	LOH	LOH	LOH	Y
3	R	LOH	LOH	Y
4	LOH	LOH	LOH	Y
5	R	R	R	N
6	R	R	R	N
7	R	R	R	N
8	R	R	R	N
9	R	R	R	N
10	LOH	LOH	LOH	Y
11	LOH	LOH	LOH	Y
12	R	R	R	N
13	LOH	LOH	LOH	Y
14	LOH	LOH	LOH	Y
15	LOH	LOH	LOH	Y
16	LOH	R	R	Y
17	LOH	R	R	Y
18	LOH	LOH	LOH	Y
19	LOH	LOH	LOH	Y
20	LOH	R	R	Y

LOH, loss of heterozygosity; R, retention of alleles.

LOH inpatient tumor: N, no; Y, yes.

Other potential mechanisms of downregulation of FABP7 could be because of methylation of the FABP7 promoter region. However, CpG islands of the FABP7 promoter region have not been reported, and in the evaluation of the promoter region near the FABP7 gene open-reading frame, no obvious CpG islands were found.

FABP7 mRNA detection in blood as a surrogate CTC biomarker

We examined FABP7 expression as a surrogate mRNA biomarker for CTCs in cells in the blood of melanoma patients. An optimal direct CTC qRT assay for FABP7 detection was established using blood from melanoma patients and normal healthy donors. The direct CTC qRT assay in melanoma has been used for other melanoma CTC biomarkers (Koyanagi et al., 2005b). Normal donor peripheral blood lymphocytes (PBLs) were shown to be negative for the optimal FABP7 qRT assay. A melanoma cell spiking experiment, performed to assess qRT assay sensitivity, showed a direct correlation between absolute FABP7 mRNA copies and the number of melanoma cells added to normal PBLs (Figure 6). FABP7 mRNA was positive for 10 melanoma cells mixed with 10⁷ PBLs in three separate experiments. These studies indicated the feasibility and sensitivity of the FABP7 qRT assay to detect CTCs in blood.

Figure 6. Melanoma cell spiking of PBLs.

FABP7 qRT quantification of serially diluted melanoma cells (10³ 10², 10¹, 10, and 0 cells) mixed with 10⁷ normal PBLs. A mean of absolute mRNA copy numbers are given. Small bars represent SD. This is a representative of three assays.

FABP7 mRNA was measured by an optimized qRT assay in 134 blood specimens from melanoma patients of different AJCC stages, and 93 blood specimens from healthy donors (Figure 7). Absolute mRNA copies of FABP7 in melanoma patients' blood varied in levels, and in healthy donors there was no detection of FABP7 mRNA.

Figure 7. FABP7/GAPDH mRNA copy ratios in blood from healthy donors and melanoma patients.

Horizontal bars denote mean mRNA copies. P<0.0001.

To determine the potential of FABP7 expression as a surrogate CTC biomarker of early-stage disease, blood was assessed from different AJCC stage melanoma patients. FABP7 was positive in blood from 17 of 72 (24%) patients with stage I/II disease, 11 of 34 (32%) patients with stage III disease, and 6 of 28 (21%) patients with stage IV disease. Stage I/II, stage III, or stage IV were significantly (P<0.001) greater than normal donor blood (0 of 93), respectively. As expected, CTC detection in advanced stage melanoma blood did not increase likely due to the loss of FABP7 gene expression. Previously, we have shown CTC detection by known melanoma biomarkers in advanced stage melanoma blood to increase significantly to early-stage disease (Koyanagi et al., 2005a, b). CTC detection in early-stage disease was less frequent, as expected, since metastasis is less frequent. However, early-stage melanomas (stage II/III) after primary tumor removal can reoccur as systemic metastases within 10 years of follow-up (Balch et al., 2001).

DISCUSSION

In this study, we assessed the expression of FABP7 in melanoma progression assessing different stages of primary and metastatic melanomas, and normal human skin as a control. Interestingly, the FABP7 mRNA expression in 69% primary melanoma tissues and 19% metastatic melanoma tissues FABP7 mRNA copies was significantly above normal human skin controls. IHC analysis of paired primary and metastatic melanomas was confirmed in FABP7 protein. Our study analysis on a well-defined large cohort of primary and metastatic melanomas of different AJCC stages showed that, as melanomas progressed, FABP7 expression decreased significantly. This observation suggested that FABP7 expression was highly related to melanoma metastasis development.

It has been shown recently that tumor expression of FABP7 is correlated with poor survival outcome in patients with glioblastoma (Liang et al., 2005). We previously showed that FABP7 is involved in the proliferation and invasion of melanoma cells (Goto et al., 2006). This suggests that FABP7 may be involved in primary tumor development. In this study, we have shown that FABP7 does not appear to be involved in primary tumor progression. No significant changes in progression for stage I to III primary melanomas were detected. We show that FABP7 expression in metastases of melanoma was associated with a significantly lower relapse-free and poorer overall survival expression. The findings were similar to the results reported with glioblastoma patients (Liang et al., 2005). These findings of the gioblastoma and our study strongly suggest that FABP7 expression in tumors has a significant role in disease outcome. Interestingly, the loss of expression may be related to genomic instability of the chromosome 6q23 region.

The studies showed that LOH on the 6q21–23 region containing the FABP7 gene is associated with and can explain to some extent the downregulation or loss of FABP7 expression in metastatic cutaneous melanoma. Deletions and genomic instability of the long arm of chromosome 6 (6q) are among the most common chromosomal alterations in cutaneous melanoma (Millikin et al., 1991; Fujiwara et al., 1999; Shirasaki et al., 2001; Stark and Hayward, 2007). Previously, we have shown that LOH is present in the 6q22–23 region in melanoma tissues (Fujiwara et al., 1999). As cutaneous melanoma progresses, genomic instability becomes more prominent and LOH becomes more frequent in specific chromosome regions (Taback et al., 2004).

We found LOH in 50% of metastatic melanomas but none in primary melanomas. This is a significant genomic aberration that has not been previously shown in detail for other LOH analyses on melanomas. LOH was detected in 50% of metastatic melanoma specimens examined, whereas FABP7 mRNA was detected in only 19% of metastatic melanomas. Loss of the chromosome region may be a genomic instability consequence that favors downregulation and loss of FABP7 expression leading to better prognosis. Interestingly, in glioblastoma tumors, LOH on chromosome 6q21–23 is frequent and associated with poor prognosis (Wooten et al., 1999; Wong et al., 2006). Early invasive melanomas in FABP7 may support melanoma growth. Other forms of generic regulation of FABP7 could occur. Epigenetic mechanisms were not likely to be the major cause of downregulation of FABP7 expression; CpG islands in the immediate promoter region of the FABP7 gene were not detectable.

Detection of CTCs are promising prognostic blood biomarkers for the detection of recurrence or progression of melanoma (Koyanagi et al., 2005a, b; Mocellin et al., 2006). We hypothesized that FABP7 may be used to detect CTCs and subclinical metastasis in blood in early stages of melanoma. It is known that subclinical distant metastasis occurs at early stages of melanoma such as in AJCC stage II/III disease; however, clinical detection (systemic recurrence) is not evident until after extensive follow-up time beyond the primary tumor removal (Morton et al., 2006). Identification of early-stage melanoma cells released from metastatic tumors may be of clinical utility in identifying these high-risk melanoma patients who have potential of developing systemic metastatic disease. It has been shown that the death rates of patients who have AJCC stage I/II disease after primary tumor removal within 12 years can range from about 5–50%, depending on the disease status and how early the disease was detected (Balch et al., 2001; Morton et al., 2006). Detection of FABP7 biomarker as a CTC surrogate in the blood indicates its potential utility for detection of biomarker subclinical metastasis in early-stage disease, whereas the value of FABP7 in stage IV disease is unfortunately limited because of the loss of gene expression in metastatic tumors. The results of the CTC analysis of FABP7 also support the findings of FABP7 expression in primary and metastatic tumors. In addition, the detection of FABP7 in CTCs may be associated with poor prognosis, and be useful as a predictive biomarker to identify high-risk patients. We could not assess the relation between the detection of FABP7 in CTCs and the prognosis of melanoma patients in this study, because patients have not been followed up long enough.

In summary, FABP7 expression is significantly decreased because of LOH on the 6q21–23 region containing FABP7 gene in metastases of melanoma. FABP7 expression in metastasis of melanoma is associated with a significantly poor disease outcome. These findings suggest that FABP7 may function as a tumor progression gene of melanoma, and can be a valuable tumor prognostic biomarker. FABP7 can be used as a potential diagnostic of early-stage melanoma systemic CTC spreading in the blood.

MATERIALS AND METHODS

Cell lines, tissues, and blood

Ten human metastatic melanoma cell lines established and characterized at the John Wayne Cancer Institute (JWCI) were used: ME-19, ME-25, ME-26, ME-27, ME-28, ME-29, ME-30, ME-31, ME-32, and ME-33. All cell lines were grown in GIBCO RPMI 1640 (Invitrogen, Carlsbad, CA) medium supplemented with 10% heat-inactivated fetal bovine serum. Cells were cultured at 37 °C in a humidified atmosphere containing 5% CO₂, as previously described (Mori et al., 2005).

Paraffin-embedded archival tissue (PEAT) specimens from AJCC melanoma patients were used for comparative analysis: 30 stage I primary melanomas, 25 stage II patients, and 32 stage III patients, and metastatic melanomas from 32 stage III patients, and 36 stage IV patients. PEAT specimens (n=74) of 37 paired primary and metastatic melanomas were used for IHC analysis. Thirteen PEAT specimens of normal skin from sun-exposed anatomical sites of these patients were used as controls. All PEAT specimens were histopathology verified by a pathologist.

Blood specimens were obtained preoperatively from 134 patients who underwent surgical treatment of AJCC stage I/II (N=72), stage III (N=34), and stage IV (N=28) melanoma. Blood was collected in 2×4.5ml sodium citrate tubes, and the first several ml were discarded to eliminate skin-plug contamination. Nucleated cells from patient blood were processed to RNA, quantified, and assessed for CTCs by qRT as previously described (Koyanagi et al., 2005b). All blood specimens were then coded by a computer-generated number so that the qRT–PCR study could be conducted in a blinded manner. Peripheral blood lymphocytes from 93 healthy normal donors served as controls. Total cells in the blood were collected by using Purescript RBC lysis solution (Gentra, Minneapolis, MN) following the manufacturer's instructions. Informed human subject were approved by Saint John's Health Center/JWCI institutional review board, and the approval of the medical ethics committee of the Shinshu U. School of Medicine was obtained for all patient specimens. The study was conducted according to the Declaration of Helsinki Principles and all patients provided signed, written informed consent to participate in this study.

RNA isolation, primers, and probes

Total RNA was extracted from cell lines and PBL specimens using Tri-Reagent (Molecular Research Center Inc., Cincinnati, OH), as previously described (Takeuchi et al., 2004; Koyanagi et al., 2005b). For RNA extraction from tissue specimens, 5×10 μm-thick sections were cut from each PEAT block with a sterile microtome blade and placed in sterile microcentrifuge tubes (Eppendorf, Westbury, NY). After deparaffinization, specimens were treated with a proteinase K digestion buffer for 3 hours before RNA extraction, as previously described (Goto et al., 2006). Total RNA was extracted, isolated, and purified using a modified RNAWiz (Applied Biosystems, Foster City, CA) phenol-chloroform extraction method, as previously described (Goto et al., 2006; Kim et al., 2006). RNA from cell lines and specimens was quantified and assessed for purity by UV spectrophotometry and a RiboGreen detection assay (Invitrogen), as previously described (Koyanagi et al., 2005b).

The development and optimization strategy for primer and probe sequences of the qRT assay was previously described elsewhere (Takeuchi et al., 2004). Fluorescence resonance energy transfer probe sequences for the FABP7 PCR product were designed to enhance the specificity of the assay. Specific primers were designed to sequence at least one exon-exon region. The FABP7 primer sequences were as follows: 5′-AAGTCTGTTGTTAGCCTGGA-3′ (forward); 5′-AGGGTCATAACCATTTTGC-3′ (reverse); 5′-FAM-TAC AGAAATGGGATGGCAAAGAAA-BHQ-1-3′ (fluorescence resonance energy transfer). The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) primer sequences were as follows: 5′-GGGTGTGA ACCATGAGAAGT-3′ (forward); 5′-GACTGTGGTCATGAGTCCT-3′ (reverse); 5′-FAM-CAGCAATGCCTCCTGCACCACCAA-BHQ-1-3′ (fluorescence resonance energy transfer). Expression of the housekeeping gene GAPDH served as an internal reference for mRNA integrity.

qRT assay

For cell lines and PBL specimens, reverse transcription (RT) reactions were performed using Moloney murine leukemia virus RT (Promega, Madison, WI) with oligo-dT primer as previously described (Koyanagi et al., 2006). Each PCR was subjected to 40 cycles of denaturation at 95 °C for 60seconds, annealing at 56 °C for 60 seconds, and extension at 72 °C for 60 seconds for FABP7; and 40 cycles of denaturation at 95 °C for 60 seconds, annealing at 55 °C for 60 seconds, and extension at 72 °C for 60 seconds for GAPDH. For PEAT specimens, RT reactions were performed with both oligodT and random hexamers (Takeuchi et al., 2004). The qRT assay was performed with the iCycler iQ RealTime PCR Detection System (Bio-Rad Laboratories, Hercules, CA) using 250 ng of total RNA for each reaction, and the mRNA copy number was calculated.

Each sample was assayed in duplicate with appropriate positive (melanoma cell line) and negative controls (normal PBLs) and reagent controls. Specific plasmid controls of FABP7 and GAPDH were synthesized as described previously (Goto et al., 2006), and standard curves for each gene were generated with a threshold cycle of six serial dilutions of plasmid templates (10⁶-10¹ copies). PCR efficiency evaluated from the slopes of the curves was between 95 and 100%. The correlation coefficient for all standard curves was ≥0.99. The product size of FABP7 and GAPDH was confirmed by gel electrophoresis, and then optimized for assay conditions for qRT. The cutoff for FABP7 positivity was set at 0.0011 (ratio of FABP7/GAPDH mRNA copies), a value that exceeded the mean ratio of the FABP7/GAPDH mRNA copies plus 3 SD above normal human skin controls.

DNA extraction

Genomic DNA was extracted from melanoma lines and autologous PBLs, as previously described (Spugnardi et al., 2003). Tissue sections were cut from 14 primary and 20 PEAT metastases, stained with hematoxylin and eosin, and microdissected as previously described (Shinozaki et al., 2004). Normal tissue (normal skin for primary melanoma and lymph node metastases comparison, and normal liver for liver metastases comparison) taken from the same slide was the source of control DNA (Fujimoto et al., 2004). Dissected tissues were digested with proteinase K overnight at 50 °C, followed by heat denaturation for 10 minutes at 95 °C. Lysates were directly used as templates for PCR, as previously described (Fujimoto et al., 2004).

Microsatellite analysis

Loss of heterozygosity (LOH) was assessed using three known microsatellite markers (D6S268, D6S1702, and D6S262) encompassing the FABP7 gene locus at 6q22.31. LOH was assessed as previously described for microsatellite instability (Fujimoto et al., 2004). The primer sets were as follows: D6S268, 5′-CTAGGTGG CAGAGCAACATA-3′ (forward) and 5′-AAAAGGAGGTCATTT TAATCG-3′ (reverse); D6S1702, 5′-AAGGAGCCATTTT TGTGC-3′ (forward) and 5′-TGCCAGCATTTTTTGGA-3′ (reverse); D6S262, 5′-ATTCTTACTGCTGGAAAACCAT-3′ (forward) and 5′-GGAGCATA GTTACCCTTA AAATC-3′ (reverse). Forward primers were labeled with WellRED dye-labeled phosphoramidites (Beckman Coulter Inc.) as previously described (Fujimoto et al., 2004). The PCR amplification was performed in a 10-μl reaction volume with 1-μl template for 40 cycles of 30 seconds at 94 °C, 30 seconds at 56 °C, and 30 seconds at 72 °C, followed by a 7-minute final extension at 72 °C. PCR product separation was performed using capillary array electrophoresis (CAE CEQ 8000XL; Beckman Coulter Inc., Fullerton, CA). Peak signal intensity and relative size were generated by fragment analysis system software (Beckman Coulter Inc.) (Fujimoto et al., 2004). Tumors were scored as LOH when one allele showed 50% or more reduction of peak intensity for tumor DNA when compared with the corresponding allele identified in the control DNA, or retention of alleles (R) when the allele showed less than 50% reduction of peak intensity for tumor DNA as previously described (Fujimoto et al., 2004). Autologous patients' PBLs served as normal controls for respective melanoma cell line analyses.

IHC staining

Immunohistochemical staining of melanoma cell lines with polyclonal rabbit anti-FABP7 antibodies was performed as previously described (Goto et al., 2006). All IHC staining was assessed by two independent dermatopathologists.

Statistical analysis

Receiver operating characteristic curve analyses was conducted to evaluate the diagnostic accuracy of FABP7, using primary melanoma/normal skin tissue as a binary outcome and the FABP7 value as possible cutoff points for computation of sensitivity and specificity. Area under the curve was computed to measure the predictive power. A Wilcoxon rank sum test was used to assess differences in FABP7 mRNA expression between primary melanomas and normal skin, between primary melanomas and metastatic melanomas, and between blood specimens from melanoma patients and healthy volunteers. Fisher's exact test was used to assess the frequency of FABP7 expression in primary versus metastatic melanomas, the frequency of LOH in primary versus metastatic melanomas, and the association between the expression of FABP7 and AJCC stage in CTC blood. McNemar's test was used to compare the frequency of FABP7 expression in the paired samples of primary and metastatic melanomas. Relapse-free survival and overall survival from surgery were used for outcome measurement. Survival curves were generated using the Kaplan-Meier method. Univariate analysis of relapse-free survival and overall survival for FABP7 detection and clinicopathologic factors (gender, age, primary site, pT stage, Clark level, N classification, M classification, and AJCC stage) were carried out using the log-rank test. Cox proportional hazard model was developed for multivariate analysis to examine the association of FABP7 detection with relapse-free survival and overall survival controlling for clinicopathologic variables. A stepwise procedure was used for prognostic variable selection. SAS software (SAS Institute, Cary, NC) was used for statistical analysis and all tests were two-sided with a significance level of P<0.05.

Abbreviations

Ab

antibody

AJCC

American Joint Committee on Cancer

BLBP

brain lipid-binding protein

CAE

capillary array electrophoresis

CTC

circulating tumor cells

FABP7

fatty acid-binding protein 7

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

IHC

immunohistochemistry

LOH

loss of heterozygosity

PBL

peripheral blood lymphocytes

PEAT

paraffin-embedded archival tissue

qRT–PCR

quantitative tealtime polymerase chain reaction

RT

reverse transcription