Excessive TNF-α Production Results from Reduction of c-Fos Via Aberrant Expression of Mir-34a in Neutrophils from Myelodysplastic Syndrome Patients

The molecular basis of ineffective hematopoiesis in myelodysplastic syndromes (MDS) has yet to be defined. Elevation of tumor necrosis factor (TNF)-α in bone marrow plasma has been considered to be a cause of apoptosis of hematopoietic progenitor cells, and this could result in ineffective hematopoiesis in MDS. However, the mechanism of TNF-α elevation has not been elucidated. We recently found aberrant regulation of c-fos mRNA stability in neutrophils isolated from MDS patients (Feng et al, PLoS ONE, 2013). Since the expression levels of RNA-binding proteins that regulated c-fos mRNA stability were not altered in the majority of the patients and no mutations were detected in c-fos mRNA, microRNAs were possibly involved in the insufficient stabilization of c-fos mRNA.

In this study, we chose 20 microRNAs targeting c-fos according to four different databases and compared their expression levels in neutrophils between MDS patients and healthy volunteers. Among them, miR-34a and miR-155 were significantly increased in MDS (p<0.05, p<0.05). Aberrantly high expressions of miR-34a and miR-155, which were defined as greater levels than the average in the healthy cells plus 2 standard deviations, were respectively detected in 14 and 13 out of 23 patients with RCUD, RCMD, and RAEB-1. We next examined the effects of the microRNA overexpression on c-fos expression. When miR-34a and miR-155 were introduced into HL60 cells by electroporation to increase their levels to 100-fold or higher than those in the control cells, c-fos protein levels decreased to 35.0 ± 10.9% and 47.8 ± 6.8% of that in the control cells, respectively, while mRNA was not altered. In neutrophils from the patients, c-fos protein but not mRNA was significantly decreased in 13 out of 17 patients tested. The c-fos protein levels were inversely correlated with miR-34a levels (r = -0.618, p<0.05) but not with those of miR-155 (r = -0.135), suggesting that c-fos protein levels were more affected by miR-34a than miR-155.

Recently, it was reported that c-fos physically interacted with nuclear factor κB (NF-κB) p65 to inhibit transcription of TNF-α in response to pro-inflammatory cytokines such as lipopolysaccharide (LPS) in monocytic cells (Koga et al, Immunity, 2009). Therefore, we hypothesized that neutrophils with reduced c-fos would overproduce TNF-α in response to inflammatory stimuli in MDS. To test our hypothesis, siRNA for c-fos was introduced into HL60 cells that were differentiated toward a neutrophil-like phenotype by 1.25% DMSO. The treatment with siRNA decreased c-fos protein levels to 58.8 ± 19.6-fold of that in the control cells. When stimulated with 1 μM LPS for 3 hours, the increase rates of TNF-α mRNA were greater in c-fos siRNA-treated cells (41.2 ± 25.7-fold, p<0.05) than in the control cells (6.1 ± 2.3-fold). Chromatin immunoprecipitation assay demonstrated that LPS increased the binding of NF-κB p65 protein to the promoter region of TNF-α DNA by 2.7 ± 1.0-fold in the control cells. In c-fos siRNA-treated cells, the amounts of NF-κB p65 bound to the TNF-α DNA promoter without stimulation were 2.1 ± 0.3-fold greater than those in the unstimulated control cells, which was further increased to 5.6±0.3-fold by stimulation with LPS (p<0.05 compared to LPS-stimulated controls). These data suggested that neutrophils, as well as monocytic cells, transcribed greater amounts of TNF-α when c-fos was decreased. Neutrophils from patients with similar c-fos levels to those in the controls (Group A, n = 5), from the patients with reduced c-fos (Group B, n = 5), and from healthy controls (n = 8) were cultured in the presence of 1 μM LPS for 3 hours. TNF-α mRNA in control neutrophils was elevated 14.9 ± 3.2-fold. The increase in Group A was 10.1 ± 4.5-fold with no significant difference from the controls, while that in Group B was greater (31.5 ± 15.1-fold, p<0.05). Although neutrophils from Group A secreted similar amounts of TNF-α into the culture medium (143.5 ± 65.7 pg/mL) to the controls (150.6 ± 91.5 pg/mL), Group B produced significantly greater amounts of TNF-α (735.4 ± 65.7 pg/mL, p<0.05 vs. controls, p<0.05 vs. Group A). Thus, our data suggest that the reduction of c-fos resulting from aberrant expression of microRNAs contributes to overproduction of TNF-α under inflammatory stimuli in MDS.