Abstract
Recently, several novel mechanisms of tumor progression were characterized in chronic lymphocytic leukemia (CLL), stressing an important role of tumor microenvironment. Direct cell-cell interactions as well as soluble factors might support tumor growth. Pro-survival signals might be supported by the tumor necrosis factor (TNF) or some distinctive members of TNF family as well as angiogenic factors. Thalidomide has been shown to be a promising immunomodulatory drug (IMiD) that targets not only leukemic cells but also the tumor microenvironment and by inhibiting angiogenesis. However, so far little is known about the molecular effects of IMiDs in-vivo. Therefore, we investigated cellular and molecular changes induced by a 7 day thalidomide mono-treatment of CLL patients. In forty cases TNF expression levels were assessed using a high sensitivity ELISA test, and T-cell subpopulations were analyzed by flow cytometry. To evaluate the influence of thalidomide on gene expression (GEP), 20 paired pre-treatment and thalidomide-treated samples were analyzed using Affymetrix U133plus2.0 microarrays. Thalidomide therapy was effective in decreasing the number of CLL cells (CD5+CD19+) from 51.75 G/L to 31.7 G/L after treatment. The number of CD3 lymphocytes showed no significant change during thalidomide therapy. No effect on CD4+ as well as CD8+ T cells was observed. Interestingly, we found significant decrease in the number of CD4+CD25hiFOXP3+ T regulatory cells (Tregs) after thalidomide (p=0.05). Thalidomide therapy did not reduce the number of other T-cell subpopulations reported to possess regulatory properties such as CD8+CD28−, CD4+GITR+, CD4+CD62L+ and CD3+TCRγδ+. The changes in Tregs during thalidomide therapy differed in cases who responded to thalidomide with a WBC reduction compared to non-responders. In cases with a WBC reduction there was a greater fold of Tregs reduction. With regard to TNF, we observed no significant changes in the TNF plasma levels after thalidomide treatment. However, the expression of TNF R1 was significantly higher in patients without WBC reduction following thalidomide (p=0.03). Gene expression profiling revealed a thalidomide-induced signature comprising 123 differentially expressed genes (paired t-test with random variance model, p<0.001). Upon thalidomide treatment we observed an up-regulation of genes known to be involved in mediating thalidomide response, such as FAS and CDKN1A. In addition, we detected novel candidate genes, such as STAT1 and IKZF1. By gene set enrichment analysis investigating 271 BioCarta pathways we found 22 pathways to be significantly deregulated (random variance model, p<0.005). These included pathways involved in apoptosis, angiogenesis, cytokine, PDGF and p38 MAPK signaling. Cases with a WBC reduction following thalidomide administration were characterized by a lower expression of pro-survival cytokines such as IL8 and lower levels of TGFB1, whereas genes involved in apoptosis like e.g. CASP1 were more highly expressed than in non-responders. On the other hand, non-responders showed higher ZAP70 expression as determined by microarray analysis and higher expression levels of anti-apoptotic genes, such as TRAF1, and genes involved in angiogenesis, such as ECGF1. However, only thalidomide responders showed also a correlation with CD40L signaling. We were also able to define a thalidomide response signature. For example, pretreatment samples derived from cases responding to thalidomide with a WBC reduction differed from non-responders with regard to the expression of JUN and CASP9. In conclusion, our study provides novel biological insights into the molecular effects of thalidomide, which might act (i) by enhancing apoptosis of CLL cells and (ii) by reducing Tregs, thereby enabling T-cell dependent tumor rejection.
A random variance model for detection of differential gene expression in small microarray experiments
