Differential microRNA profiles and their functional implications in different immunogenetic subsets of chronic lymphocytic leukemia

Abstract

Critical processes of B-cell physiology, including immune signaling through the B-cell receptor (BcR) and/or Toll-like receptors (TLRs), are targeted by microRNAs. With this in mind and also given the important role of BcR and TLR signaling and microRNAs in chronic lymphocytic leukemia (CLL), we investigated whether microRNAs could be implicated in shaping the behavior of CLL clones with distinct BcR and TLR molecular and functional profiles. To this end, we examined 79 CLL cases for the expression of 33 microRNAs, selected on the following criteria: (a) deregulated in CLL versus normal B-cells; (b) differentially expressed in CLL subgroups with distinct clinicobiological features; and, (c) if meeting (a) + (b), having predicted targets in the immune signaling pathways. Significant upregulation of miR-150, miR-29c, miR-143 and miR-223 and downregulation of miR-15a was found in mutated versus unmutated CLL, with miR-15a showing the highest fold difference. Comparison of two major subsets with distinct stereotyped BcRs and signaling signatures, namely subset 1 [IGHV1/5/7-IGKV1(D)-39, unmutated, bad prognosis] versus subset 4 [IGHV4-34/IGKV2-30, mutated, good prognosis] revealed differences in the expression of miR-150, miR-29b, miR-29c and miR-101, all down-regulated in subset 1. We were also able to link these distinct microRNA profiles with cellular phenotypes, importantly showing that, in subset 1, miR-101 downregulation is associated with overexpression of the enhancer of zeste homolog 2 (EZH2) protein, which has been associated with clinical aggressiveness in other B-cell lymphomas. In conclusion, specific miRNAs differentially expressed among CLL subgroups with distinct BcR and/or TLR signaling may modulate the biological and clinical behavior of the CLL clones.

Figure 1

Differential miRNA expression in M-CLL versus UM-CLL. Only miRNAs with fold difference in expression of ≥1.50 and significantly different expression in M-CLL versus U-CLL (p ≤ 0.05) are represented in the graphs. Among them, miR-15a, downregulated in M-CLL, has the highest fold difference. The y axis depicts the relative expression (2^−ΔCt). Error bars represent standard deviation (SD). *p ≤ 0.05 compared with U-CLL.

Figure 2

Differentially expressed miRNAs in subset 1 versus 4 CLL cases. Only miRNAs with fold difference in expression of ≥1.50 and significantly different expression (p ≤ 0.05) are represented in the graphs. Among them, miR-101, downregulated in subset 1, shows the highest fold difference. The y axis depicts the relative expression (2^−ΔCt). Error bars represent SD. *p ≤ 0.05 compared with subset 4.

Figure 3

miR-15a/MAPK8 and miR-101/FOS are negatively correlated at mRNA but not at protein level. MAPK8 (JNK) is the target with the highest prediction score for miR-15a. (A, C) Correlation and linear regression between miR-15a/MAPK8 and miR-101/FOS mRNA levels, respectively. ΔCt values were used and the correlation is statistically significant (p < 0.01). The slope in both cases is negative in both cases, which means that when miRNA expression is upregulated, mRNA expression is downregulated. (B, D) Protein levels of MAPK8 and FOS in representative M/U-CLL and subset 1 and 4 cases, respectively. MAPK8 is overexpressed in U-CLL. (E) miR-101 forced overexpression in one subset 1 case by using miR-101 mimic induced FOS protein downregulation compared with untransfected cells or cells transfected with negative control. (F) Transfection of subset 4 cells with miR-101 inhibitor induced FOS downregulation, yet to a lesser extent compared with miR-101–forced overexpression.

Figure 4

miR-101 differentially regulates EZH2 at mRNA and protein levels in subset 1 versus 4 CLL cases. (A) EZH2 mRNA levels in subsets 1 and 4 (four and four cases, respectively). The y axis represents relative expression (2^−ΔCt). The box-and-whisker plots repesent median with minimum and maximum values. EZH2 is upregulated in subset 1 cases (fold difference in expression of 2.3; p ≤ 0.05). (B) Correlation and linear regression between miR-101 and EZH2 mRNA levels. ΔCt values were used and the correlation is statistically significant (p ≤ 0.05). The slope is negative, which means that when miR-101 expression is upregulated, EZH2 expression is downregulated (r = −0.75). (C) Protein levels of EZH2 in subsets 1 and 4 (six and eight cases, respectively). EZH2 protein levels are higher in subset 1 (p < 0.005). (D) Correlation and linear regression between miR-101 relative expression (2^−ΔCt) and EZH2 protein levels (r = −0.7, p < 0.05). (E) In subset 1 CLL cells, miR-101 mimic induces downregulation of EZH2 protein levels compared with untransfected cells and cells transfected with negative control. (F) In subset 4 CLL cells, miR-101 inhibitor induces EZH2 protein upregulation compared with untransfected cells or cells transfected with negative control.