Inhibition of miR-21 restores RANKL/OPG ratio in multiple myeloma-derived bone marrow stromal cells and impairs the resorbing activity of mature osteoclasts

Abstract

miR-21 is an oncogenic microRNA (miRNA) with an emerging role as therapeutic target in human malignancies, including multiple myeloma (MM). Here we investigated whether miR-21 is involved in MM-related bone disease (BD). We found that miR-21 expression is dramatically enhanced, while osteoprotegerin (OPG) is strongly reduced, in bone marrow stromal cells (BMSCs) adherent to MM cells. On this basis, we validated the 3'UTR of OPG mRNA as miR-21 target. Constitutive miR-21 inhibition in lentiviral-transduced BMSCs adherent to MM cells restored OPG expression and secretion. Interestingly, miR-21 inhibition reduced RANKL production by BMSCs. Overexpression of protein inhibitor of activated STAT3 (PIAS3), which is a direct and validated target of miR-21, antagonized STAT3-mediated RANKL gene activation. Finally, we demonstrate that constitutive expression of miR-21 inhibitors in BMSCs restores RANKL/OPG balance and dramatically impairs the resorbing activity of mature osteoclasts. Taken together, our data provide proof-of-concept that miR-21 overexpression within MM-microenviroment plays a crucial role in bone resorption/apposition balance, supporting the design of innovative miR-21 inhibition-based strategies for MM-related BD.

Keywords: OPG; RANKL; miR-21; miRNAs; multiple myeloma bone disease.

Figure 1. miR-21 upregulation in HS-5 correlates with OPG downregulation

A. Quantitative RT-PCR analysis of miR-21 and OPG expression in HS-5 cultured alone (HS-5 alone) or adherent to either MM cell lines (HS-5 + RPMI 8226; HS-5 + U266) or primary MM cells (HS-5 + MM PCs) and exposed to healthy PBMCs (HS-5 + Healthy PBMCs). miR-21 expression increased by 6, 0-fold and 3, 46-fold in RPMI 8226 - HS-5 co-culture (p < 0.05), by 3, 9-fold and 6, 25-fold in U266 – HS-5 co-culture (p < 0.05) and by 2, 8-fold and by more than 8-fold (p < 0.05) in primary MM cells – HS-5 co-culture after 24 and 48 hours respectively. OPG expression significantly decreases in the presence of highest miR-21 expression levels (p < 0.05). Mean of Ct values were normalized to RNU44 housekeeping snoRNA or GAPDH and expressed as 2-DDCt value calculated using the comparative cross threshold method. Values represent mean ± SD of three independent experiments. B. ELISA analysis of OPG secretion in HS-5 cultured alone or co-cultured with RPMI 8226 or Primary MM cells. OPG concentration was reported as fold expression and each value, expressed in pmol/l, was normalized to HS-5 alone. Values represent the mean ± SD from three independent experiments. * indicates p < 0.05.

Figure 2. miR-21 is upregulated and OPG downregulated in MM patient-derived BMSCs

Quantitative RT-PCR analysis of miR-21 and OPG from BMSCs of MM patients and from BMSCs of healthy donors after 3 weeks of culture period. Mean of Ct values were normalized to RNU44 housekeeping snoRNA or GAPDH and expressed as 2-DCt value calculated using the comparative cross threshold method. miR-21 expression levels of MM patients vs healthy donors: p = 0.016; OPG expression levels of MM patients vs healthy donors: p = 0.018.

Figure 3. OPG expression is directly controlled by miR-21

A. Quantitative RT-PCR analysis of miR-21 and OPG expression in HS-5 transfected with synthetic miR-21 (miR-21 HS-5) or miRNA scrambled (miR-NC HS-5). Mean of Ct values were normalized to RNU44 housekeeping snoRNA or GAPDH and expressed as 2-DDCt value calculated using the comparative cross threshold method. Values represent mean ± SD of three independent experiments. B. ELISA analysis of OPG secretion calculated as fold expression. Each value expressed in pmol/l was normalized to control (miR-NC HS-5). Values represent mean ± SD of three independent experiments. C. Dual-luciferase assay of HS-5 co-transfected with firefly luciferase constructs containing the 3′UTR of OPG or its mutant lacking miR-21 target sequence (nts 601–675) and miR-21 or miRNA scrambled. The data are shown as relative luciferase activity of miR-21-transfected cells as compared with the control (miR-NC) of three independent transfections and were normalized to renilla luciferase activity. * indicates p < 0.05.

Figure 4. Constitutive miR-21 inhibition significantly increases OPG production

A, upper panels. Immunoblot detection of OPG in anti-miR-21 and control vector HS-5 cultured for 48 and 72 hours with RPMI 8226, U266 and primary MM cells. γTubulin was used as loading control. B, lower panels. ELISA analysis of OPG secretion in RPMI 8226 – HS-5 co-culture, U266 – HS-5 co-culture, Primary MM cells – HS-5 co-culture. OPG concentration was reported as fold expression and each value, expressed in pmol/l, was normalized to control (control vector HS-5). Values represent the mean ± SD from three independent experiments. * indicates p < 0.05.

Figure 5. Constitutive miR-21 inhibition reduces RANKL production

A, B and C left panels. Immunoblot detection of mRANKL (transmembrane RANKL) and sRANKL (soluble RANKL) in anti-miR-21 and control vector HS-5 cultured for 48 and 72 hours with RPMI 8226 (A), U266 (B) and primary MM cells (C) γTubulin was used as loading control for RPMI 8226, GAPDH for U266 and primary MM cells – HS-5 co-cultures. A, B and C right panels. ELISA analysis of RANKL secretion in RPMI 8226 (A), U266 (B) and primary MM cells (C) – HS-5 co-cultures. RANKL concentration was calculated as fold expression and each value, expressed in pmol/l, was normalized to control. Values represent mean ± SD of three independent experiments. * indicates p < 0.05.

Figure 6. miR-21 is upregulated in primary patient BMSCs adherent to MM cells and its constitutive inhibition restores RANKL/OPG ratio

A. Quantitative RT-PCR analysis of miR-21 and OPG expression in primary patient BMSCs alone (MM BMSCs alone) and adherent to either MM cell lines (MM BMSCs + RPMI 8226; MM BMSCs + U266). miR-21 expression increased by 7-fold and 9, 5-fold in RPMI 8226 - MM BMSCs co-culture (p < 0.05), 3, 1-fold and 5, 6-fold in U266 - MM BMSCs co-culture (p < 0.05) after 24 and 48 hours respectively. OPG expression significantly decreases in the presence of highest miR-21 expression levels (p < 0.05). Mean of Ct values were normalized to RNU44 housekeeping snoRNA or GAPDH and expressed as 2-DDCt value calculated using the comparative cross threshold method. Values represent mean ± SD of three independent experiments. B. ELISA analysis of OPG secretion in primary patient BMSCs cultured alone or with RPMI 8226. OPG concentration was reported as fold expression and each value, expressed in pmol/l, was normalized to MM BMSCs alone. Values represent the mean ± SD from three independent experiments. C and D, left panels. Immunoblot detection of OPG in anti-miR-21 and control vector MM BMSCs cultured for 48 and 72 hours with RPMI 8226 (C) and U266 (D) GAPDH was used as loading control. C and D central panels. Immunoblot detection of mRANKL (transmembrane RANKL) and sRANKL (soluble RANKL) in anti-miR-21 and control vector MM BMSCs cultured for 48 h and 72 h with RPMI 8226 (C) and U266 (D) GAPDH was used as loading control. C and D, right panels. ELISA analysis of OPG and RANKL concentration expressed in RANKL/OPG ratio in RPMI 8226 (C) and U266 (D) – MM BMSCs co-cultures. Each value of RANKL (pmol/l) was divided to each value of OPG (pmol/l). Values represent mean ± SD of three independent experiments. *indicates p < 0.05.

Figure 7. Descriptive cartoon of IL-6, miR-21 and STAT3 axis

A. IL-6 signaling leads to RANKL and miR-21 expression in BMSCs through STAT3 activation. miR-21 overexpression inhibits OPG production and, in a positive feedback loop, promotes RANKL gene activation by reducing PIAS3, specific inhibitor of active STAT3 (pSTAT3). The RANKL/OPG imbalance in MM microenviroment results in a severe perturbation of bone homeostasis. B. Inhibition of miR-21 restores PIAS3 and OPG expression: PIAS3 interferes with RANKL gene expression by inhibiting STAT3 activation, while in MM microenviroment OPG antagonizes RANKL binding on this receptor RANK, thus restoring the physiologic OCLs activity.

Figure 8. Inhibition of miR-21 suppresses RANKL through PIAS3 upregulation

A. Immunoblot detection of PIAS3 (upper panel) and pSTAT3 (lower panel) in anti-miR-21 and control vector HS-5 cultured with RPMI 8226. GAPDH and γTubulin were used as loading control. B. Immunoblot detection of pSTAT3 (upper panel) and mRANKL (transmembrane RANKL) and sRANKL (soluble RANKL) (lower panel) in anti-miR-21 and control vector HS-5 transfected with specific PIAS3 or scrambled stealth siRNAs and cultured with RPMI 8226. γTubulin was used as loading control.

Figure 9. Exposure to co-culture medium from HS-5 stably expressing miR-21 inhibitors suppresses the OCLs activity

A, left panel. Analysis of lacunae generation on dentin surfaces of OCLs cultured for 4–5 days with serial dilutions of culture medium (100% CM, 50% CM and 25% CM) form RPMI 8226 – control vector or anti-miR-21 - HS-5 co-culture (OCLs + CM control vector HS-5; OCLs + CM anti-miR-21 HS-5). Red circles showed on representative dentin surfaces indicate some single bone resorption lacunae identified by blue toluidine staining. A, right panel, B and C. Analysis of lacunae generation by NIH imageJ software that considered each pit as a dark area on dentin surface. Results shown are the mean from lacunae number in three dentin fields out of three independent experiments. *indicates p < 0.05.