Overexpression of miR-125a in myelodysplastic syndrome CD34+ cells modulates NF-κB activation and enhances erythroid differentiation arrest

Abstract

Myelodysplastic syndromes (MDS) are characterized by impaired proliferation and differentiation of hematopoietic stem cells. The participation of toll-like receptor (TLR)-mediated signaling in MDS is well documented. Increased TLR signaling leads to the constitutive activation of NF-κB, which mediates inflammation, cell proliferation and apoptosis. In addition, the TLR pathway induces the expression of miRNAs which participate in the fine-tuning of the inflammatory response. miRNAs also regulate other biological processes, including hematopoiesis. miR-125a and miR-125b are known modulators of hematopoiesis and are abnormally expressed in several hematologic malignancies. However, little is known about their role in MDS. NF-κB-activating ability has been described for both miRNAs. We studied the role of miR-125a/miR-125b in MDS and their relationship with TLR signaling and hematopoietic differentiation. Our results indicate that miR-125a is significantly overexpressed in MDS patients and correlates negatively with patient survival. Expression of miR-99b, which is clustered with miR-125a, is also directly correlated with prognosis of MDS. Both miR-125a and miR-99b activated NF-κB in vitro; however, we observed a negative correlation between miR-99b expression and the levels of TLR2, TLR7 and two downstream genes, suggesting that NF-κB activation by the miRNA cluster occurs in the absence of TLR signaling. We also show that TLR7 is negatively correlated with patient survival in MDS. In addition, our data suggest that miR-125a may act as an NF-κB inhibitor upon TLR stimulation. These results indicate that miR-125a is involved in the fine-tuning of NF-κB activity and that its effects may depend on the status of the TLR pathway. Furthermore, we observed that miR-125a inhibits erythroid differentiation in leukemia and MDS cell lines. Therefore, this miRNA could serve as a prognostic marker and a potential therapeutic target in MDS.

Figure 1. Expression of miR-125a and miR-125b in BM CD34⁺ cells.

(A–B) Relative expression of miR-125a and miR-125b in CD34⁺ cells from MDS patients (N = 48/47, respectively) and healthy donors. In (A), one data point (Y = 386.31) is outside the axis limits. Statistical significance versus healthy donors: **P<0.01. (C) Correlation between relative expression of miR-125a and miR-125b in MDS CD34⁺ cells. Six data points are outside the axis limits.

Figure 2. Correlation between the relative expression of miR-125a and overall survival in MDS.

Levels of miR-125a are inversely correlated with patient survival.

Figure 3. Expression of miR-99b and let-7e in BM CD34⁺ cells.

(A–B) Relative expression of miR-99b and let-7e in CD34⁺ cells from MDS patients (N = 48/45, respectively) and healthy donors. One outlier removed by Grubb's method (α = 0.05) in each case. (C–D) Correlation between the relative expression of miR-125a and miR-99b or let-7e, respectively, in MDS CD34⁺ cells. In (C) and (D), three data points are outside the axis limits.

Figure 4. Correlation between the relative expression of miR-99b and overall survival in MDS.

Levels of miR-99b are inversely correlated with patient survival.

Figure 5. Effect of ectopic expression of miR-125a and/or miR-99b on NF-κB activity.

NF-κB activation was determined after 48 hours from transfection of Meg-01 cells with miRNA mimics and luciferase vectors. Results are relative to cells transfected with mock RNA and expressed as mean ± SEM of n = 3 independent experiments. Statistical significance: *P<0.05; ***P<0.001.

Figure 6. Relationship between the miR-99b/let-7e/miR-125a cluster members and genes from the TLR2-NF-κB pathway in MDS.

“Low” and “high” expression cohorts were established based on comparison with the mean relative expression value. (A–B) Correlation between the relative expression of miR-99b and TLR2 or MyD88, respectively, in MDS CD34⁺ cells. In (B), one outlier removed by ROUT method. (C–E) Correlation between the relative expression of miR-125b and TLR2, MyD88 or JMJD3, respectively in MDS CD34⁺ cells. In (E), five outliers removed by ROUT method. (F) Correlation between relative expression of miR-125a and JMJD3. Seven outliers removed by ROUT method. Statistical significance: *P<0.05; ***P<0.001.

Figure 7. Expression of TLR7 in MDS.

(A) Relative expression of TLR7 in BM CD34⁺ cells. mRNA levels of TLR7 in CD34⁺ cells from MDS patients (N = 42) and healthy donors were measured by qPCR. (B–E) Correlation between the relative expression of TLR7 and TLR2, miR-99b, miR-125b or miR-125a, respectively, in MDS CD34⁺ cells. “Low” and “high” expression cohorts were established based on comparison with the mean relative expression value. In (B), two data points are outside the axis limits; in (C), three outliers removed by ROUT method; in (D), four outliers removed by ROUT method; in (E), seven outliers removed by ROUT method. Statistical significance: *P<0.05; ***P<0.001.

Figure 8. Correlation between the relative expression of TLR7 and overall survival in MDS.

Levels of TLR7 are directly correlated with patient survival.

Figure 9. Effect of TLR stimulation and miR-125a inhibition on NF-κB activity in KG1 cells.

NF-κB activation was measured after 24 hours from nucleofection of KG1 cells with the luciferase vectors and treatment with ASO. Results are expressed as relative to cells transfected with mock and represent mean ± SEM of n = 4 independent experiments. Method disclosure: technical problems regarding the endogenous Renilla control were experienced during these luciferase assays; only one experiment out of four efficiently expressed the Renilla luciferase and could be properly normalized. Because normalized results were almost identical to non-normalized data, we conducted a joint statistical analysis of the four experiments. Statistical significance: ***P<0.001.

Figure 10. Effect of miR-125a inhibition on Ara-C-stimulated erythroid differentiation of K562 cells.

K562 cells pre-treated with 1 µM ASO for 6 hours and treated with ASO for 48 hours were plated in MC for 4 days before being counted and collected for the corresponding assays. (A) Effect on cell density after 48 hours (n = 5). (B) Colony formation assay (n = 5). (C) Benzidine-positive colony count (n = 3). The proportion of erythroid-like colonies after 4 days from plating is expressed as the number of benzidine-positive colonies, normalized to the total colony number. (D) Expression of CD71 (TFRC) in K562 colonies. Statistical analysis represents a grouped analysis of the main effect of ASO (n = 4). (A–D) Data represent mean ± SEM. Statistical significance: *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

Figure 11. Characterization of MDS-L cells.

(A) Basal expression levels of miR-125a and miR-125b in myeloid leukemia and MDS cell lines. miRNA levels are represented as the relative value to expression in K562 cells. Expression of miR-125a in MDS-L cells was 5–10 times higher than that of miR-125b (note that relative levels of miR-125a were calculated independently from those of miR-125b). Data represent mean ± SEM of n = 4 independent experiments. (B) Spontaneous differentiation of MDS-L cells after a 7-day culture in methylcellulose, expressed as the relative expression of differentiation markers, assessed by qPCR in MDS-L colonies collected after 7 days of incubation in methylcellulose. Data represent mean ± SEM of n = 3 independent experiments. Statistical significance: *P<0.05.

Figure 12. Effect of the inhibition of miR-125a and TLR2-NF-κB pathway in MDS-L cells.

Black bars represent cells treated with 5 µM control peptide, and striped bars represent cells treated with 5 µM MyD88-inhibitor peptide. (A) Effectiveness of miR-125a ASO in MDS-L cells, expressed as relative miR-125a expression levels after 48 hours of treatment with 1 µM ASO. (B) Cell density after 48 hours of treatment. (C) Colony formation assay after 7 days from plating in methylcellulose. (D–F) Relative expression of EPO-R, GYPA and CD71 (TFRC), respectively, measured by qPCR after a 7-day methylcellulose culture of cells previously treated for 48 hours with 1 µM ASOs. (A–F) Data represent mean ± SEM of n = 3 independent experiments. Statistical significance: *P<0.05.

Figure 13. Proposed model for miR-125a regulation in MDS cells.

Discontinuous arrows represent speculations and unknown mechanisms; consecutive arrows represent known pathways that do not need to be explained for the understanding of the figure. (A) In the absence of TLR signaling, the miR-99b/let-7e/miR-125a cluster (and very likely also miR-125b) is coordinately upregulated, via transcriptional activation by either NF-κB as a part of a positive feedback loop or by other transcription factors (TF). miR-125a, presumably in collaboration with miR-99b, enhances NF-κB activation, probably through the inhibition of the TNF-induced NF-κB inhibitor TNFAIP3 and/or other inhibitors, such as IκBε. Thus, the expression of the miR-99b/let-7e/miR-125a cluster may favor survival of hematopoietic cells, protecting them from the deadly effects of TNF-α. (B) Upon TLR signaling, NF-κB is activated through a cascade of adaptor proteins. Under these conditions, the miR-99b/let-7e/miR-125a cluster is not expressed (or expressed at low levels) but the expression of miR-125a might be independently induced by unknown (maybe disease-related) mechanisms, such as the initiation of transcription at alternative promoter regions or, more likely, the differential processing of the primary transcript in one or more stages of miRNA biogenesis. Highly increased levels of miR-125a could preferentially target the mRNA of one or more genes downstream of TLRs and participate in the negative modulation of proinflammatory signaling. In MDS, elevated miR-125a levels in cells with normal TLR signaling (A) would favor sustained NF-κB activation and pro-survival effects; while in those cells with increased TLR/MyD88 levels and activation (B), high miR-125a levels would negatively modulate NF-κB activation. However, in this case, the hyperactivation of the TLR/MyD88/NF-κB pathway would probably mask the inhibitory effects of miR-125a. Additionally, in both cases, (A) and (B), overexpression of miR-125a in MDS patients would favor aberrant differentiation so its effects in either case would be detrimental for the course of the disease.