Arpc1b Gene Is a Candidate Prediction Marker for Choroidal Malignant Melanomas Sensitive to Radiotherapy

Abstract

purpose. Choroidal malignant melanomas (CMMs) are the most common primary intraocular tumors in adult humans. Although radiotherapy is commonly used to treat the melanomas, the therapeutic effects are unpredictable. The purpose of this study was to search for a gene(s) that can predict the success of radiotherapy for CMMs.

methods. The cell lines 92-1, OCM-1, and OMM-1 were established from patients with CMM, and radiation sensitivity was determined using the colony-formation assay. RNA was extracted from nonirradiated cells, and gene expression analysis was performed using a microarray containing 10,800 genes. The up- or downregulated genes were verified by real-time PCR using other cancerous cell lines in which radiation sensitivity had been documented.

results. Analysis of radiation survival curves showed that cell line 92-1 was radiation sensitive and OCM-1 and OMM-1 lines were radiation resistant. The results of microarray analyses showed that 34 genes were differentially expressed in the OCM-1 and OMM-1 cell lines compared with the 92-1 cell line. The validity of the expression level of 13 of the 34 genes that were identified by microarray was confirmed by PCR. From the analysis of the different radio-sensitivity cancer cell lines, the Arpc1b gene was selected as a prediction marker gene for sensitivity of CMM to radiotherapy.

conclusions. Gene expression analysis of CMM cell lines can be used to search for radiation sensitivity prediction markers. Comprehensive gene expression profiles of radiation-sensitive and/or resistant cell lines may provide new insights into the mechanisms of resistance or sensitivity to radiation therapy.

Choroidal malignant melanomas (CMMs) are the most common primary intraocular tumors in human adults ¹ and are the only potentially fatal ocular malignancy. However, the incidence of this type of tumor is quite low. ² The mode of treatment has changed with clinical experience, but for large tumors and for those with complications, enucleation remains the treatment of choice. ³ The Collaborative Ocular Melanoma Study (COMS) reported that there was no increase in survival time by pre-enucleation radiation for large tumors. ⁴ However, for medium-sized tumors, radiation therapy remains the treatment of choice. ⁵ The mortality rates, however, have been found not to differ between iodine-125 brachytherapy and enucleation over a 12-year survival period. Gragoudas et al. ⁶ reported that proton beam therapy and Char et al. ⁷ that helium ion therapy can reduce local recurrences but have increased anterior segment complications. In addition, the effects of treatment are unpredictable, so that determining the indications for radiotherapy is an important aspect in the treatment of malignant melanomas and in reducing side effects. 

Microarray technology permits a comprehensive analysis of the genes expressed in cells and tissues and is widely used for the investigation of cancerous cells. ⁸ It is a valuable tool for understanding the basic biology of malignant tumors and has been a great help in the treatment of cancerous cells. ⁹ There have been many microarray analyses on different types of malignant tumors, but none for CMM. 

The purpose of this study was to search for genes in CMM cells that would predict whether the cells were sensitive to radiotherapy. To accomplish this, we compared the expression profiles of radio-sensitive and radio-resistant CMM cell lines by using an oligonucleotide microarray containing 10,800 probes. 

Materials and Methods

Cell Lines

The CMM cell line 92-1 ¹⁰ was kindly provided by Martine J. Jager and H. Monique H. Hurks (Leiden University Medical Center, Leiden, The Netherlands). The OMM-1 ¹¹ cell line was provided by Gré P. M. Luyten (Rotterdam University Hospital, Rotterdam, The Netherlands) and the OCM-1 ¹² line by June Kan-Mitchell (Department of Immunology and Microbiology, Wayne State University School of Medicine, Detroit, MI). Five esophageal carcinoma cell lines, TE1, TE2, TE3, TE10, and TE13, were obtained from the Cell Resource Center for Biomedical Research Institute of Development, Aging, and Cancer, Tohoku University, and three esophageal carcinoma cell lines—YES2, YES5, YES6—were obtained from Yamaguchi University. The other esophageal carcinoma cell line, TTn, was established at Chiba University. The lung carcinoma cell lines A549, H1299, SB3, and SBC5 were obtained from the American Type Culture Collection (ATCC, Manassas, VA). 

Radiation Sensitivity

The surviving fraction was measured using a standard colony-formation assay. Cells were irradiated with different radiation doses (2, 4, and 6 Gy) from a linear accelerator (Al 1-mm filter, 2.20 Gy/min; MBR-1520R-3; Hitachi, Tokyo, Japan) and seeded into 60-mm dishes. Colonies obtained after 10 to 14 days were stained with cresyl violet and counted in routine manner. 

Microarray Analysis

Two pairs of cell lines (92-1-OCM-1; 92-1-OMM-1) that were sensitive and resistant to ionizing radiation were selected for the gene microarray analysis (AceGene Human Oligo Chip; Hitachi Software Engineering). We have analyzed gene expression profiles using an oligonucleotide microarray containing 10,800 probes. The total RNA was extracted from each CMM cell line (TRIzol Reagent; Invitrogen, Carlsbad, CA), according to the manufacturer’s instructions. A fluorescence probe was synthesized by incorporating the modified nucleotide, Cy3- or Cy5- aminoallyl UTP, into the aRNA during in vitro transcription (Amino Allyl Message Amp aRNA kit; Ambion, Austin, TX). After purification, the dye-labeled aRNA was used for microarray hybridization. Each microarray analysis experiment was performed twice. 

Validation of Microarray Analysis by RT-PCR

To verify the microarray data, reverse transcription (RT)-PCR was performed. The cDNA templates were synthesized from 1 μg of the total RNA with reverse transcriptase (SuperScript II; Invitrogen, Carlsbad, CA) and oligo-dT primers. The PCR program was set for 3 minutes at 95°C for 1 cycle followed by 95°C for 20 seconds and 62°C for 1 minute for 30 cycles and a final extension at 4°C on a PCR system (Gene Amp 9700; Applied Biosystems Inc. [ABI], Foster City, CA). 

We selected 13 candidate genes to predict the radiation sensitivity and used real-time PCR to quantify the expression level of each of the 16 cell lines with known surviving fractions. The real-time PCR program was 10 minutes at 95°C for 1 cycle followed by 95°C for 15 seconds and 63°C for 1 minute for 40 cycles on a real-time PCR system (model 7300; ABI). Primer sequences of the 13 genes are shown in Table 1 . 

The 16 cell lines studied were 3 CMM cell lines, 9 esophageal cancer cell lines, and 4 lung cancer cell lines. β-Actin was used as a control to calculate the gene expression score. 

Statistical Analyses

Pearson’s correlation coefficient and Fisher’s exact test was used to analyze the significance of the association of results of real-time RT-PCR and surviving fraction at 2.0 Gy irradiation (SF2). Pearson’s correlation coefficient was calculated (StatView; Hulinks, Berkeley, CA) and the correlation coefficient (r) and significance level (P) were determined. The Fisher exact test was performed with a cutoff score: gene expression = 2.5 and SF2 = 0.6. 

Results

Cells from the 92-1, OCM-1, and OMM-1 cell lines were irradiated, and survival curves were determined from the postirradiated cells by colony formation assay. The cells were then classified as resistant or sensitive according to the colony formation assay. Cell line 92-1 was classified as radiation-sensitive, and cell lines OCM-1 and OMM-1 were classified as radiation resistant (Fig. 1) . 

After the cell lines were classified, oligonucleotide microarray analysis was used to determine different levels of gene expression of nonirradiated 92-1/OCM-1 and 92-1/OMM-1 cells. Microarray analysis revealed 21 genes of the resistant cell line that had elevated expression levels greater than twofold and 13 genes of the sensitive cell line that expressed levels of more than twofold. Thirteen of the 34 genes selected as candidate gene predictors of radiosensitivity by gene expression analysis were confirmed by RT-PCR (Fig. 2)and are listed in Table 2 . 

Other cancerous cell lines were also treated with ionizing radiation and used to determine the genetic factors that were common to the cancerous cells. We used 13 cancerous cell lines (9 cell lines of esophageal cancer, 2 cell lines of small cell lung cancer, and 2 cell lines of pulmonary adenocarcinoma) to investigate the validity of our hypothesis (Table 3) . A wide range of sensitivity to radiation was detected among the different cell types, and differences in radiation sensitivity were also seen within the same cell type. The SF2 of 16 cell lines were determined to be between 0.130 and 0.883. 

We selected the 13 genes that were confirmed by RT-PCR and determined their expression levels in the 16 cell lines: 3 CMM lines and the 13 cell lines determined by real-time RT-PCR. The correlation between gene expression and SF2 was determined using Pearson’s correlation coefficient and the Fisher exact test on the 13 genes. The results of the analysis (Table 4)showed that only the Arpc1b gene was statistically significant by both Pearson’s correlation coefficient (r = 0.592, P = 0.0142) and the Fisher exact test (P = 0.041). The quantitative expression levels of Arpc1b gene for each of the 16 cell lines determined by real time RT-PCR are shown in Figure 3 . As for the Arpc1b gene expression, a high tendency was recognized in resistant cells, and a low tendency was recognized in sensitive cells. A scatterplot with the regression line showing the relationship between Arpc1b gene expression and SF2 for each of the 16 cell lines is shown in Figure 4 . The expression level of Arpc1b correlated positively with the SF2 values. 

From these findings, we designated the Arpc1b gene as a potential marker for radio-resistant cells. However, the remaining 12 genes showed no statistically significant correlation between SF2 and gene expression status in this study. 

Discussion

Radiotherapy is an important treatment modality for esophageal squamous cell cancer, ¹³ pulmonary adenocarcinoma, small-cell lung cancer, ¹⁴ and many other types of cancerous cells. Radiotherapy and concomitant chemotherapy combined with surgery shows promising results in cases of unresectable esophageal cancers. ¹⁵ The energy particles of ionizing radiation can damage cellular DNA, either by direct interaction with the DNA or through the creation of chemically reactive free radicals. If the energy of the particle is high enough, such damage can lead to cell death. ¹⁶  

Double-strand breaks (DSBs) are a common type of DNA damage produced by ionizing radiation. Two major types of DSB repairs, nonhomologous end-joining (NHEJ) and homologous recombination (HR), are present to repair and maintain the genomic structure in mammalian cells. Many proteins participate in these pathways, and deficiencies in the repair of proteins sometimes influence the radiosensitivity. Therefore, the proteins that play a role in repairing the DSBs most likely determine the radiosensitivity of the cells. ¹⁷  

We used microarray analysis to investigate genes that could be predictors of the radiosensitivity in cancerous cells. However, the proteins involved in DNA damage repair pathways were not detected by microarray analysis, probably because the three CMM cell lines were used without irradiation, and these proteins were not upregulated. The Arpc1b gene was correlated with the radiosensitivity (SF2) in the CMM cells and other cancerous cells studied. At present, there are no reports that suggest an association of the expression level of the Arpc1b gene and radiosensitivity. Therefore, we can only speculate on its function in cells with high resistance to ionized radiation. 

Our earlier results showed that even after irradiation, there were no repair genes with increased expression when nonirradiated OCM-1 and irradiated OCM-1 (2 Gy, 4 hours after irradiation) were compared by microarray analysis (data not shown). The irradiation power may have to be higher to detect differences in gene expression when performing microarray analysis on cells before and after irradiation. 

The Arpc1b gene is designated as an actin-related protein 2/3 complex, subunit 1b (arpc1b; 41 kDa). It is a subunit of the actin-related protein 2/3 complex (Arp 2/3 complex), which is implicated in the control of actin polymerization in cells. The Arp 2/3 complex is composed of seven subunits: Arp2, Arp3, p16-Arc, p20-Arc, p21-Arc, p34-Arc, and p41-Arc (Arpc1b). ¹⁸  

When considering the characteristics of malignant tumor cells, cell motility, invasion, and subsequent metastasis are important properties. ¹⁹ The Arp 2/3 complex regulates actin filament cross-linking, which could selectively stabilize actin filaments to assemble into microspikes in lamellipodia, thus enabling cell movement. ²⁰ The Arp 2/3 complex may nucleate the formation of new actin filaments in the cell, ²⁰ and an increase in Arp2/3 expression and subsequent cell motility may contribute to the radioresistance of the cells by progression of metastasis. 

We have not found any reports on the association of anticancer therapy and the Arp2/3 complex, thus making our study novel. Some studies have examined gene expressions by microarray using irradiated cancer cells, but Arp2/3 was not identified in any of these studies. ^{21 22 23} The direct influence of Arpc1b in metastasis, DNA repair, or protection from ionizing radiation is yet to be investigated. In contrast, there are reports of a decrease of all seven of the subunits of the Arp2/3 complex in gastric cancerous cells. ²⁴ It was concluded that the decrease in the Arp2/3 complex results in a decrease in actin polymerization that will then contribute to dysplastic cell morphology, a characteristic of cancerous cells. The difference in results from our results may be due to differences in the characteristics of the cancer cell lines. Investigations into the association of increased radioresistance due to the expression of the other subunits of the Arp2/3 complex are needed. 

There was no statistically significant correlation between gene expression and SF2 in the other 12 genes. The reason for this is unclear, but the use of cell lines may have some influence on the outcome of our experiment. The use of clinical samples in the future may reveal new results and shed light on the association of radiosensitivity and expression of some of these genes. 

Clinical samples of CMM are very difficult to obtain in ophthalmology because biopsy is not usually performed in such cases. However, we will continue to search for new samples to investigate the expression of the Arpc1b gene and its role in radiosensitivity using clinical samples obtained from other branches of medicine. 

 Submitted for publication June 26, 2005; revised January 9, 2006; accepted April 25, 2006.

 Disclosure: K. Kumagai, None; Y. Nimura, None; A. Mizota, None; N. Miyahara, None; M. Aoki, None; Y. Furusawa, None; M. Takiguchi, None; S. Yamamoto, None; N. Seki, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Primer Sequences of the 13 Selected Genes

Symbol	Primer Sequence (F)	Primer Sequence (R)
Bcl2a1	5′-aggtccaagcaaaacgtcca-3′	R-ggggcaatttgctgtcgtag3′
Igfbp7	5′-acaacctggccattcagacc3′	R-tagctcggcaccttcacctt3′
Gage7	5′-gaaggggaaccagcaactca3′	R-gcgttttcacctcctctgga3′
Gpx1	5′-accagtttgggcatcaggag3′	R-gagcttggggtcggtcataa3′
Lgals3	5′-ccacgcttcaatgagaacaac3′	R-acccgatgattgtactgcaac3′
Arpc1b	F-atctgggatgtgaagagcttg3′	R-atgaccagcatagtgctttgg3′
Sat	F-tctaagccaggttgcaatgag3′	R-caacgccactggtaataaagc3′
Mgc4365	F-tcgaggccctagattttcttc3′	R-tgctgtttgaaggtctctgtg3′
Cdc16	F-aaaggttcttcagccaagctc3′	R-gatgcgttctgaggaatcaac3′
Tyrp1	F-tgataccctgggaacactttg3′	R-ggaataaaaaggaggcgtgtc3′
A2m	F-caaggtggatttgagcttcag3′	R-tcttcatcgtcctggtcattc3′
Serpinf1	F-cttcaaggggcagtgggtaa3′	R-gtcactttcaggggcaggaa3′
CtsL	F-tgaggcaacagaagaatcctg3′	R-ctgtgctttcaaatccgtagc3′

 F, forward; R, reverse.

Figure 1.

 

SFs of three CMM cell lines that were calculated by colony-formation assay. According to the SF2 (surviving fraction at 2 Gy irradiation) score, a definition of radiosensitivity as <0.4 and radioresistance as >0.6 was adopted. (▴) OMM-1; (▪) OCM-1; and (□) 92-1.

Figure 2.

 

Thirteen of 34 genes selected as candidate predictor genes indicating sensitivity to radiotherapy of choroidal malignant melanomas by oligonucleotide microarray analysis were confirmed by RT-PCR. GAPDH was used as a control. Lane 1: 92-1; lane 2: OCM-1; and lane 3: OMM-1.

Table 2.

 

Candidate Genes that Were Detected by Oligonucleotide Microarray Analysis

A. Upregulated Genes in Radiation-Resistant Cell Lines
Symbol	Accession No.	Gene	Ratio of Expression (x-fold)
						OCM/92-1	OMM/92-1
Bcl2al	NM-004049	Bcl2-related protein 21	7.81	8.23
Igfbp7	NM-001553	Insulin-like growth factor binding protein 7	6.90	9.88
Gage7	NM-021123	G antigen7	4.00	7.50
Gpx1	NM-000581	Glutathion peroxidase1	3.48	2.64
Lgals3	NM-002306	lectin, galactoside-binding, soluble, 3	3.42	6.24
Arpc1b	NM-005720	Actin related protein2/3 complex 41kd	3.07	5.82
Sat	NM-002970	Spermidine/spermine n1 acetyltransferase	2.97	4.17
Mgc4365	NM-024330	Hypothetical protein	2.76	2.48
Cdc16	NM-003903	Cdc 16 homolog	2.63	3.35

B. Downregulated Genes in Radiation Resistant Cell Lines
Symbol	Accession No.	Gene	Ratio of Expression (x-fold)
						92-1/OCM	92-1/OMM
Tyrp1	NM-000550	Tyrosinase-related protein 1	12.13	2.93
A2m	NM-000014	Alpha 2 macroglobulin precursor	11.67	8.15
Serpinf1	NM-002615	Serine proteinase inhibitor	9.65	5.86
Cts1	NM-001912	Cathepsin 1	3.83	3.08

 The microarray analyses were performed twice to obtain the Cy5/Cy3 ratio between 92-1 and OCM-1, 92-1 and OMM-1. The significant ratio is estimated more than 2-fold increase in expression. (A) More than two-fold ratio of expression for both OCM-1 and OMM-1 compared with 92-1. (B) More than two-fold ratio of expression for 92-1 compared with both OCM-1 and OMM-1.

Table 3.

 

The 16 Cell Lines Used for the Real-Time RT-PCR

Cell Line	SF2	Cell Origin
92-1	0.130	Choroidal malignant melanoma
TE2	0.165	Esophageal squamous cell carcinoma
YES5	0.236	Esophageal squamous cell carcinoma
SBC3	0.237	Small-cell lung carcinoma
YES2	0.573	Esophageal squamous cell carcinoma
TE3	0.571	Esophageal squamous cell carcinoma
TE10	0.576	Esophageal squamous cell carcinoma
TTn	0.629	Esophageal squamous cell carcinoma
TE13	0.633	Esophageal squamous cell carcinoma
TE1	0.636	Esophageal squamous cell carcinoma
YES6	0.669	Esophageal squamous cell carcinoma
OCM	0.670	Choroidal malignant melanoma
A549	0.733	Pulmonary adenocarcinoma
OMM	0.790	Choroidal malignant melanoma
H1299	0.780	Pulmonary adenocarcinoma
SBC5	0.883	Small-cell lung carcinoma

Table 4.

 

Pearson’s Correlation Coefficient and Fisher Exact Test Results

Gene	Pearson’s Correlation Coefficient			Fisher Exact Test Significance Level (P)
		Correlation Coefficient (r)	Significance Level (P)
Bcl2a1	0.302	0.262	0.475
Igfbp7	0.196	0.474	0.145
Gage7	0.472	0.064	0.475
Gpx1	0.083	0.765	0.358
Lgals3	0.301	0.264	0.145
Arpc1b	0.592	0.014	0.041
Sat	0.313	0.244	0.696
Mgc4365	0.289	0.284	0.687
Cdc16	0.202	0.460	0.182
Tyrp1	−0.410	0.116	0.242
A2m	−0.485	0.056	0.242
Serpinf1	−0.470	0.066	0.242
CtsL	0.019	0.945	0.833

 Correlation coefficient (r) and significance level (P) are shown for 13 distinct genes confirmed by RT-PCR. Only the arpc1b gene was statistically significant when calculated by both Pearson’s correlation coefficient (r = 0.592, P = 0.0142) and Fisher exact test (P = 0.041). Arpc1b data are indicated in bold. Cutoff score of the Fisher exact test was gene expression = 2.5 and SF2 = 0.6.

Figure 3.

 

Real-time RT-PCR was performed on 16 cell lines for 13 genes confirmed by using RT-PCR. The quantitative expression level of the Arpc1b gene is shown for each of the 16 cell lines. The expression scale on the y-axis was calculated by dividing the Arpc1b expression data by the expression of β-actin. Error bars, SD of results in four independent experiments.

Figure 4.

 

Scatterplot showing the regression of arpc1b expression in cell lines on radiosensitivity (SF2) calculated by real-time RT-PCR data. The correlation between gene expression and SF2 was statistically significant (Pearson’s correlation coefficient, r = 0.592; P = 0.0142). The expression shown on the y-axis was calculated by dividing the arpc1b expression data by the expression of β-actin. The horizontal axis represents SF2 of the 16 cell lines.