Abstract
Background: Older age has been shown to consistently correlate with inferior survival in diffuse large B-cell lymphoma (DLBCL). This is likely attributable in part to poorer performance status and inability to tolerate therapy, however, potential molecular differences associated with age are not well defined. In addition, the impact of sex has been shown to be important in DLBCL. Models that identify adverse tumor biology related to sex remain largely unexplored in DLBCL. Via a comprehensive systems biology approach (e.g., Yang Y, Nature Commun. 2014; Kuchenbaecker KB, Nature Genet. 2015), we performed detailed global transcriptome analyses from the Cancer Genome Atlas (TCGA) to investigate the biologic dynamics from pre-treatment/baseline DLBCL based on age and sex. Furthermore, a novel unbiased method was used to identify "key genes" and related signaling networks most strongly associated with adverse DLBCL biology.
Methods: From TCGA research network (cancergenome.nih.gov), lymphoid neoplasm DLBCL mRNA level 3 data type with a total of 48 data sets from untreated DLBCL patients (pt) was available with 33 data sets containing relevant age and sex. Since median age was 58 years for this data set, we defined older pts as ≥58 years (vs ≤57 for younger patients). Significant genes were determined by T-test with p-value <0.05 for all comparisons to take to pathway analysis. Pathway analysis of selected genes were performed using fold change > ±1.2 comparing old to young pts and observing pathway relationships using Ingenuity Pathway Analysis (IPA) software. Upstream regulator and biofunction analysis was done with IPA. Gene Set Enrichment Analysis (GSEA) was performed with FDR <0.05 for functional analysis. Furthermore, a novel unbiased systems biology method for determining key genes and association with adverse tumor biology (ie, prediction of tumor progression) were determined as previously reported (Beheshti A, Cancer Res 2015). 'Tumor progression' herein is defined as the presence of adverse tumor dynamics based on the milieu of biologic factors present (eg, tumor suppressors and oncogenes) at diagnosis.
Results: Both sex and age revealed striking, independent transcriptional differences. There were several distinct genes associated with DLBCL at older age including JUN, CR1, DNAH10, and C20orf54. Nine distinct genes were modulated by sex, regardless of age. This included XIST, which was significantly upregulated in females and DDX3Y, KDM5D, and PRKY that were downregulated in females. Furthermore, GSEA demonstrated that older age was associated with decreased metabolism and telomere functions and that globally female sex was associated with decreases in interferon signaling, transcription, cell cycle, and PD-1 signaling. In addition, through DAVID gene function classification, we discovered that the key genes for most groups strongly regulated immune function activity. Surprisingly, we also identified global downregulation of genes for older females, while older males had overall upregulation, which was in agreement with the GSEA data. Moreover, older females were predicted to have more favorable tumor dynamics vs older males as well as young females (Fig. 1). Finally, systems biology analyses revealed that JUN and CYCS signaling were the most critical inter-connected factors associated with adverse biology and likelihood of tumor progression in older and male patients, respectively (Fig. 1).
Conclusions: Collectively, these findings reinforce the importance of a detailed understanding of biology in DLBCL and the inter-connectivity of genes and signaling pathways. JUN was a predicted upstream regulator, and moreover, it was the only specific key gene that was commonly upregulated for older DLBCL patients (independent of sex). For sex, CYCS was shown to be a critically connected factor. Additionally, the majority of key genes were strongly connected to immune function activity and associated signaling networks. Altogether, understanding how molecular factors interact and change as a function of age and sex, and how this impacts tumor biology, may improve our understanding of lymphomagenesis and potentially lead to enhanced therapeutic strategies.
Figure 1. The impact of key genes on tumor biology and predicted progression in DLBCL. A) + B) Key genes illustrating the balance of tumor dynamics. C) + D) A network representation of the key genes. Figure 1. The impact of key genes on tumor biology and predicted progression in DLBCL. A) + B) Key genes illustrating the balance of tumor dynamics. C) + D) A network representation of the key genes.
Disclosures
No relevant conflicts of interest to declare.
Can Systems Biology Advance Clinical Precision Oncology?
