Abstract
Abstract 1368
The growth and accumulation of B-cell chronic lymphocytic leukemia (CLL) cells requires survival and migratory signals supplied by endogenously produced cytokines and chemokines. Multiplex cytokine and chemokine analyses revealed that CCL2, CCL3, CCL4, IL-1b, IL-2, IL-5, IL-6, IL-8, IL-10, IL-12, CXCL9, CXCL11, TNFa, IFNg, IFNa, and IL-17 levels are significantly elevated in the serum of CLL patients (84 total) as compared to healthy age-matched subjects (48 total; p<0.01 for all). Random Forest analysis was used to rank these serum factors with respect to their importance in discriminating CLL patients from normal controls. This approach revealed 8 cytokines and chemokines (CCL3, IL-10, IL-5, CCL4, CXCL11, IL-8, CCL2, and IL-17) that distinguish CLL and controls, with CCL3 being the discriminating factor with the highest weight. The Kaplan-Meier method and log-rank test were used to compare median survival in CLL patients with high (>median) versus low (<median) levels for each cytokine. This analysis revealed that high levels of CXCL10, CXCL11, and IL-1b independently correlated with poor survival (p<0.05). Unsupervised cluster analysis was performed to identify groups of cytokines and chemokines that might work in concert to modulate CLL cell biology. As expected, CLL patients and healthy subjects clustered together on different ends of the dendogram. Furthermore, CLL patients with short TFT clustered together, and expressed high levels of a discrete group of cytokines (CCL3, CCL4, CCL19, CXCL9, CXCL10, CXCL11, IFNg, IL-5, IL-10, and IL-12). We termed this group “Clus1”. Certain members of this group have previously been shown to relate to CLL pathogenesis. CCL3 and CCL4, inflammatory chemokines that regulate cell recruitment and activation, are secreted by CLL cells and trigger a cascade of events that supports the CLL survival. IL-10 and the CXCR3 ligands CXCL9, CXCL10, CXCL11, are elevated in CLL and may contribute anti-apoptotic and growth-promoting signals that promote CLL cell survival. But, our data show for the first time that this set of cytokines and chemokines may be functionally linked. Cluster analysis also revealed a group of CLL patients that had moderate levels of Clus1 in combination with a second discrete cluster of cytokines (“Clus2”; GM-CSF, IL-8, TNFa, and IL-6). Since this group of patients had a relatively long TFT, even in the presence of detectable levels of Clus1, we hypothesized that they might play a direct protective role in CLL or alternatively might serve to regulate the actions of Clus1 cytokines. Lastly, cluster analysis revealed a third group of patients that were characterized by elevated serum levels of another discrete set of cytokines and chemokines (“Clus3”; IL-1b, IL-2, IL-4, IL-15, IL-17, and IFNa). The Kaplan-Meier method and log-rank test were used to compare median survival in patients expressing higher levels of Clus1, Clus2, or Clus3 cytokines and combinations of these clusters. This analysis revealed that high serum levels of Clus1 cytokines independently correlated with shorter survival times (Clus1HI12.5 yr versus Clus1LOW >21.5 yr; p=0.01), while high serum levels of Clus3 cytokines independently correlated with longer survival times (Clus3HI 21.6 yr versus Clus3LOW 12.5 years; p=0.05). The levels of Clus2 did not correlate with CLL patient survival. Combination analyses revealed that patients with high levels of Clus1 in the absence of Clus2 had significantly shorter survival times (10 yr) relative to patients with high levels of Clus1 cytokines in the presence of Clus2 (10yr vs 16yr; p=0.0018). In contrast, there was no significant difference between survival in patients with high levels of Clus1 in either the presence or absence of Clus3HI. Taken together, this suggests a beneficial modulatory role of cluster 2 cytokines. CONCLUSIONS: Application of multiplex bead immunoassays has allowed us to identify a set of cytokines and chemokines that are elevated within the context of CLL, to rank these secreted factors with respect to their importance in discriminating CLL from controls, and to identify those factors whose expression correlates with CLL disease prognosis. Complementary bioinformatics analysis allowed identification of three distinct clusters of cytokines and chemokines whose integrated expression levels reflect different CLL patient outcomes.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal