Analysis of gene expression identifies candidate molecular markers in nasopharyngeal carcinoma using microdissection and cDNA microarray

Abstract

Purpose: Microarray analysis was used to bring a comprehensive insight into underlying molecular mechanisms and obtain a whole assessment of aberrant gene expression in nasopharyngeal carcinoma (NPC).

Methods: Combined with microdissection, gene expression profiles in 23 NPCs and 10 nontumor nasopharyngeal epithelial tissue samples were analyzed.

Results: Gene expression patterns suggested the dysregulation of the GTP/GDP-bound Ras cycle and an abnormal hyperactivity of cell cycle in NPC. Alterations in the WNT pathway suggest that this pathway may be activated in NPC. A 6-feature weighted-voting model was chosen because it represented the main characteristics of NPCs and predicted NPCs most accurately from the nontumor tissues (33 of 34 correct calls; 97.1% accuracy, Fisher's exact test, P value = 8.389 x 10(-8)).

Conclusions: The data generated in this study represent a comprehensive list of genes aberrantly regulated in NPC. The 6-feature weighted-voting model may provide an extensive list of potential molecular markers for early diagnosis.

Fig. 1

Detection of the expression of EBNA1 and LMP1 by RT-PCR in the same total RNA samples served in the microarray analysis. GAPDH was served as an internal control for each reaction

Fig. 2

Gene expression patterns of 24 NPCs (T) and 10 nontumor nasopharyngeal epithelial tissues (N), analyzed by hierarchical clustering using 213 cDNAs. Hierarchical clustering analysis was performed based on the gene expression profiles using the Euclidean distance and average linkage clustering. The specimens were divided into two subtypes based on differences in gene expression: the NPC groups versus the nontumor nasopharyngeal tissue groups. T01_S was the replicate sample of the T01_F sample

Fig. 3

Differential expression of genes in the NPC groups versus the nontumor nasopharyngeal tissue groups. Results are shown for the top 50 genes of each distinction. Each column represents a single sample, and each row represents a single gene. Gene descriptions are shown. For each gene, red indicates a high level of expression relative to the mean; blue indicates a low level of expression relative to the mean. *: The six genes were yielded by the weighted-voting class prediction; #: “T20” samples

Fig. 4

Semiquantitative RT-PCR analysis of SERPINB6, SYTL2, ADTRL1, RB1, STMN1, and DSP using the same total RNA samples in the microarray analysis. GAPDH served as an internal control for each reaction