Microarray evidence of glutaminyl cyclase gene expression in melanoma: implications for tumor antigen specific immunotherapy

Abstract

Background: In recent years encouraging progress has been made in developing vaccine treatments for cancer, particularly with melanoma. However, the overall rate of clinically significant results has remained low. The present research used microarray datasets from previous investigations to examine gene expression patterns in cancer cell lines with the goal of better understanding the tumor microenvironment.

Methods: Principal Components Analyses with Promax rotational transformations were carried out with 90 cancer cell lines from 3 microarray datasets, which had been made available on the internet as supplementary information from prior publications.

Results: In each of the analyses a well defined melanoma component was identified that contained a gene coding for the enzyme, glutaminyl cyclase, which was as highly expressed as genes from a variety of well established biomarkers for melanoma, such as MAGE-3 and MART-1, which have frequently been used in clinical trials of melanoma vaccines.

Conclusion: Since glutaminyl cyclase converts glutamine and glutamic acid into a pyroglutamic form, it may interfere with the tumor destructive process of vaccines using peptides having glutamine or glutamic acid at their N-terminals. Finding ways of inhibiting the activity of glutaminyl cyclase in the tumor microenvironment may help to increase the effectiveness of some melanoma vaccines.

Figure 1

Scree test for the Ross et al dataset. Eigenvalues from the 60 × 60 cell line correlation matrix showed that, beginning with the break in the plot after component 9, a straight scree line could be fitted to the remaining values. Such a finding suggests that the eigenvalues after component 9 represent only error variance. The term "scree" was borrowed from geology where it refers to debris (i.e. error) that has fallen to the base of a mountain. Note that for clarity of presentation, the first eigenvalue of 13.787 was not plotted.

Figure 2

Highly expressed genes within the Ross et al Melanoma Component. Gene expression levels of 5 genes within each of the 7 cell lines in the Ross et al cDNA Melanoma Component. Measurements were log ratios between the gene expression level of each cell line and a reference sample of 12 of the 60 tumor cell lines. Further details may be found in the original Ross et al publication. The means for each gene for the non-melanoma cell lines (followed by the Student t test significances for the mean difference between melanoma and non-melanoma cell lines) were as follows: QPCT -.514 (p < .001), MART1 -.055 (p = .001), TYR -.031 (P < .001), TYRP1 -.449 (P < .001), and TYRP2 -.041 (p < .001). It may be seen that the means of the non-melanoma cell lines were all very close to zero, and therefore were not plotted simply for greater clarity of presentation.

Figure 3

Scree test for the Staunton et al dataset. Eigenvalues from the 60 × 60 cell line correlation matrix showed that, beginning with the break in the plot after component 9, a straight scree line could be fitted to the remaining values. Such a finding suggests that the eigenvalues after component 9 represent only error variance. Note that for clarity of presentation, the first eigenvalue of 51.392 was not plotted.

Figure 4

Highly expressed genes within the Staunton et al Melanoma Component. Gene expression levels of 4 genes within each of the 7 cell lines in the Staunton et al Affymetrix Melanoma Component together with the mean expression levels for the combined non-melanoma cell lines. The expression values were average intensity difference units determined with Affymetrix GENECHIP software that assigned a value of 100 to all expression measurements of less than 100 units. Further details may be found in the Staunton et al publication. The Student t test significance levels for each gene, calculated between the melanoma and the non-melanoma cell lines, were as follows: QPCT (p = .033), MART1 (p = .012), MAGE1 (p = .019) and MAGE3 (p = .024). Note that for clarity of presentation, this figure used expression average intensity difference units, as they were prior to conversion to the standardized z-scores that were used in the PCA and t test calculations.

Figure 5

Scree test for the Györffy et al Affymetrix dataset. Eigenvalues from the 30 × 30 cell line correlation matrix showed that, beginning with the break in the plot after component 9, a straight scree line could be fitted to the remaining values. Such a finding suggests that the eigenvalues after component 9 represent only error variance. Note that for clarity of presentation, the first eigenvalue of 27.020 was not plotted.

Figure 6

Highly expressed genes within the Györffy et al Melanoma Component. Gene expression levels of 4 genes within each of the 5 cell lines in the Györffy et al Melanoma Component together with the mean expression levels for the combined non-melanoma cell lines. The expression values were average intensity difference units determined with Affymetrix MAS 5.0 software. Further details may be found in the Györffy et al publication. The Student t test significance levels for each gene, calculated between the melanoma and the non-melanoma cell lines, were as follows: QPCT (p < .001), MART1 (p = .010), MAGE3 (p < .001) and TYRP2 (p < .001).