Significant downregulation of platelet gene expression in metastatic lung cancer

Abstract

Platelets play a major role in the metastatic dissemination of tumor cells in vivo . Recent evidence reveals megakaryocyte-derived platelet pre-mRNA is spliced to mRNA and then translated into functional proteins in response to external stimulation. Employing a human lung cancer model, we hypothesized a subset of megakaryocyte/platelet genes exists that are significantly over or underexpressed in metastasis compared with noncancer. Microarray analysis employing platelet mRNA followed by unsupervised hierarchical clustering revealed an expression profile that includes decreased expression of 197 of the 200 platelet genes with the most altered expression (p < 1.0 × 10(-4)). Among the 608 splicing events identified between the metastasis and negative control groups, 33 highly variable genes were identified with between 3 and 13 splicing events each. In conclusion, this preliminary study reveals a platelet-based gene expression signature that differentiates metastatic lung cancer from negative controls on the basis of decreased expression of 197 of the 200 genes with the most altered expression levels. Further study may yield a prognostic tool for future metastasis among subsets of early stage lung cancer patients.

Figure 1

Principal component analysis demonstrating the variability of the gene expression data within each of the two patient groups. The fact that the points representing the five metastatic lung cancer patients (red points) are clustered distinctly apart from the seven members of the control group (blue points) is suggestive of the difference between the two groups in their respective gene expression profiles. The x, y, and z axes are the first, second, and third components that collectively capture most of the variability.

Figure 2

Gene‐level analysis: Heat map showing hierarchical clustering between metastatic lung cancer and controls. Each column is one microarray (one patient) while each differentially expressed gene is represented by one row. Gene clusters are defined by the dendrogram to the left of the heat plot. Upregulated genes are shown in red; downregulated genes are shown in green. A clear pattern of downregulation is seen in the significant majority of genes and an expression profile was derived including 200 genes with a corresponding t‐test p‐value of <1.0 × 10⁻⁴.

Figure 3

Exon‐level analysis: Example of significant under expression of a representative exon between metastatic lung cancer patients and controls. The exon number is 27 and is included within the BAZ1A gene that encodes the accessory subunit of the ATP‐dependent chromatin assembly factor (ACF). Each bar in the upper panel represents the fold change in expression for the individual probeset between the lung cancer and control samples, and the line graph designates the significance level of this difference measured in reverse log p‐value units (−log10 p‐value). The exon identified by probeset 3560768 shows markedly lower expression in the lung cancer population (−3.42 fold difference), and an elevated −log10 p‐value (11.1, equivalent to a p‐value of 0.00047). The lower panel shows that same probeset has a splicing index of −1.14 with an accompanying inverse log10 p‐value of 11.5 (P = 0.00034).