Transfer, analysis, and reversion of the fibrous dysplasia cellular phenotype in human skeletal progenitors

Abstract

Human skeletal progenitors were engineered to stably express R201C mutated, constitutively active Gs alpha using lentiviral vectors. Long-term transduced skeletal progenitors were characterized by an enhanced production of cAMP, indicating the transfer of the fundamental cellular phenotype caused by activating mutations of Gs alpha. Like skeletal progenitors isolated from natural fibrous dysplasia (FD) lesions, transduced cells could generate bone but not adipocytes or the hematopoietic microenvironment on in vivo transplantation. In vitro osteogenic differentiation was noted for the lack of mineral deposition, a blunted upregulation of osteocalcin, and enhanced upregulation of other osteogenic markers such as alkaline phosphatase (ALP) and bone sialoprotein (BSP) compared with controls. A very potent upregulation of RANKL expression was observed, which correlates with the pronounced osteoclastogenesis observed in FD lesions in vivo. Stable transduction resulted in a marked upregulation of selected phosphodiesterase (PDE) isoform mRNAs and a prominent increase in total PDE activity. This predicts an adaptive response in skeletal progenitors transduced with constitutively active, mutated Gs alpha. Indeed, like measurable cAMP levels, the differentiative responses of transduced skeletal progenitors were profoundly affected by inhibition of PDEs or lack thereof. Finally, using lentiviral vectors encoding short hairpin (sh) RNA interfering sequences, we demonstrated that selective silencing of the mutated allele is both feasible and effective in reverting the aberrant cAMP production brought about by the constitutively active Gs alpha and some of its effects on in vitro differentiation of skeletal progenitors.

Figure 1

Lentiviral transfer of Gsα^R201C mutation in hBMSCs. (A) LV vectors used for Gsα^R201C transfer. LV‐EF‐Gsα^R201C encodes the Gsα^R201C protein (HA‐tagged), controlled by the EF‐1α promoter. The bidirectional LV vector (LV‐BD‐Gsα^R201C) encodes Gsα^R201C and eGFP under the control of the PGK and CMV minimal promoters, respectively. (B) Western blot analysis of Gsα^R201C expression in LV‐EF‐Gsα^R201C‐transduced hBMSCs. Immunoblotting for HA demonstrates the specific signal for the mutated Gsα protein. Immunoblotting for Gsα demonstrates that mock‐treated and empty vector (LV‐Ctr)–transduced cells express comparable amounts of the 48‐ and 43‐kDa isoforms of Gsα (Gsα‐long and Gsα‐short, respectively) resulting from alternative splicing of exon 3. A specific increase of Gsα‐long is observed in cells transduced with LV‐EF‐Gsα^R201C. (C) Gsα^R201C expression in hBMSCs is detected by immunofluorescence, as indicated by HA immunolabeling. (D) Intracellular cAMP levels were measured in control or LV‐EF‐Gsα^R201C‐transduced hBMSCs. In the presence of IBMX (1 mM) alone or of IBMX and forskolin (10 µM), significantly higher levels of cAMP are observed in Gsα^R201C‐expressing cells compared with control cells. Data from five separate experiments in duplicate are expressed as mean ± SD. ^a p < .05 versus mock and LV‐ctr; ^b p < .01 versus mock and ctr.

Figure 2

Effect of Gsα^R201C on hBMSC proliferation and in vivo differentiation. Analysis of cell proliferation in transduced hBMSCs expressed as (A) population doublings or (B) BrdU incorporation for 2 and 24 hours. The presence of mutated Gsα^R201C does not alter BMSC proliferation. Data from three experiments are expressed as means ± SD. (C) In vivo transplantation of normal, FD‐derived (R201C) and LV‐ Gsα^R201C‐transduced hBMSCs. Note the formation of abundant bone (b) on hydroxyapatite (ha/tcp, hydroxyapatite/tricalcium phosphate), hematopoietic tissue (hem, meg = megakaryocytes), and adipocytes (ad) in control transplants (upper panels). Transplants of FD‐derived BMSCs and transduced BMSCs are recognized by the reduced amount of bone and the fibrous tissue (ft) devoid of hematopoietic cells and adipocytes filling the marrow space. Polarized light microscopy demonstrates that normal BMSCs formed lamellar bone (lb), whereas transduced BMSCs formed woven bone (wb).

Figure 3

Gsα^R201C effect on the adipogenic and osteogenic markers in undifferentiated and in vitro differentiated hBMSCs. (A) Adipogenic and osteogenic marker quantification in Gsα^R201C‐transduced hBMSCs. RT‐PCR analysis was carried out in undifferentiated uninfected and LV‐ctr‐ or LV‐Gsα^R201C‐infected hBMSCs. Significant activation of RANKL is detected in LV‐Gsα^R201C‐infected hBMSCs. (B) Adipogenic and osteogenic differentiation of Gsα^R201C‐transduced hBMSCs. hBMSCs were cultured in osteogenic and adipogenic medium, and mineralization nodules or adipocyte‐like cells were revealed with Von Kossa and oil red O staining, respectively. Both in vitro osteogenic differentiation and in vitro adipogenic differentiation of hBMSC are altered in Gsα^R201C‐transduced cells. (C) qPCR of osteogenic and adipogenic markers analyzed in differentiated hBMSCs. Abnormal differentiation was observed in Gsα^R201C‐transduced cells compared with mock‐transduced and control cells. Data from four experiments in duplicates are expressed as means ± SD. ^a p < .05 versus mock and ctr; ^b p < .01 versus mock and ctr.

Figure 4

Activity, mRNA expression, and phenotypic effects of PDEs in mutation‐transduced hBMSCs. (A) PDE activity was assayed in hBMSCs, mock‐treated or transduced with LV‐Ctr or LV‐ EF‐Gsα^R201C, using 10 µM of cAMP as a substrate, in the presence or absence of either IBMX (1 mM) or Rolipram (20 µM). ^b p < .01 versus equally treated mock and ctr cells; ^c p < .01 versus mutation‐transduced, IBMX‐ and rolipram‐treated cells. (B) qPCR analysis of the expression of PDE isoforms involved in cAMP degradation. Significant increase of expression was detected for PDE‐3b, ‐4b, ‐4d, and ‐7b isoforms in mutation‐transduced cells compared with mock‐ and LV‐Ctr‐treated cells. Values represent the mean ± SD of three experiments in duplicate. ^a p < .05 versus mock and ctr; ^b p < .01 versus mock and ctr. (C) hBMSCs were cultured in osteogenic medium supplemented with IBMX (0.5 mM) or adipogenic medium without IBMX. Mineralization nodules or adipocyte‐like cells were revealed with Von Kossa and oil red O staining, respectively. (D) PDE inhibition in osteogenic medium caused a significant upregulation of osteogenic markers in mock‐transduced and control cells at the same levels of Gsα^R201C. Adipogenesis in the absence of IMBX leads to a modest increase of adipogenic markers to the same extent in control and Gsα^R201C cells. ^a p < .05 versus mock and ctr.

Figure 5

Specific and stable LV‐mediated silencing of Gsα^R201C in normal and FD hBMSCs. (A) Schematic representation of the LV vectors expressing the Gsα RNA‐interfering sequences under the control of the H1 promoter and eGFP as a marker. (B) Design of different RNA‐interfering sequences directed to human exon 1 and human/rat Gsα^R201C exon 8. (C) Left panel: HeLa cells were first infected with the bidirectional LV‐GFP‐Gsα^R201C and subsequently superinfected with the different LV‐siGsα vectors as indicated. Specific knockdown of mutated Gsα was obtained only with LV‐siGsα^R201C‐4. Right panel: HeLa GFP‐Gsα^R201C/siCtr or HeLa GFP‐Gsα^R201C/siGsα^R201C‐4 were superinfected with LV‐EF1α‐Gsα^R201C or LV‐EF1α‐Gsα^WT. A specific downregulation of the added‐in mutated R201C Gsα was observed. In contrast, robust expression of the added‐in Gsα^WT was observed. (D) Left panel: Highly mutated in R201C, FD‐derived hBMSCs were mock infected or treated with LV‐siCtr or LV‐siGsα^R201C‐4 and after 20 days of culture, the cells were collected for Western blot analysis. A decrease in both the long and short endogenous form of Gsα was observed with LV‐siGsα^R201C‐4. Right panel: Normal hBMSCs were untreated or treated with LV‐siCtr, LV‐siGsα^R201C‐4, and LV‐siGsα human exon 1 (hu). The 20 days after infection, cells were processed for Western blotting. Both Gsα endogenous WT isoforms (long and short Gsα) were efficiently suppressed with siGsα exon 1 but remained efficiently expressed with specific siGsα^R201C‐4. (E) cAMP assay on HeLa LV‐GFP‐Gsα^R201C/siGsα^R201C‐4 demonstrates the capacity to restore the cAMP levels similar to that of the basal levels. Data are means ± SD of n = 3 experiments in duplicate. ^b p < .01 versus untransduced HeLa equally treated; ^c p < .01 versus LV‐Gsα^R201C and LV‐Gsα^R201C/LV‐siCtr cells equally treated. (F) cAMP assay on transduced FD‐derived BMSCs and normal hBMSCs. A significant reduction of cAMP levels, comparable with that recorded in normal hBMSC, was observed in FD‐LV‐siGsα^R201C‐4. ^a p < .05 of FD‐LV‐siGsα^R201C‐4 versus FD mock, FD‐LV‐siCtr, and hBMSC‐Gsα^R201C.

Figure 6

Specific lentivirus‐directed RNA interference of constitutively active Gsα restores in vitro differentiation of LV‐Gsα^R201C‐transduced skeletal progenitors. (A) Restoration of in vitro adipogenesis (top, oil red O) and mineralization (bottom, von Kossa staining) in BMSCs that were first transduced with LV‐Gsα^R201C and then superinfected with LV‐siGsα^R201C‐4. (B) Restoration of expression of mRNAs for adipogenic and osteogenic marker, qPCR analysis. Knockdown of mutated Gsα by LV‐siGsα^R201C‐4 in previously Gsα^R201C‐transduced cells reestablishes the markers expression at the same levels as seen in mock cells. siGsα^R201C‐4 does not affect the differentiation of untransduced cells. Data from four experiments in duplicate are expressed as means ± SD. ^a p < .05 versus mock, LV‐Gsα^R201C plus LV‐siGsα^R201C‐4, and LV‐siGsα^R201C‐4; ^b p < 0.01 versus mock, LV‐Gsα^R201C plus LV‐siGsα^R201C‐4 and LV‐siGsα^R201C‐4. No significant difference in expression levels occurs among mock and LV‐Gsα^R201C plus LV‐siGsα^R201C‐4, mock, and LV‐siGsα^R201C‐4 or LV‐Gsα^R201C plus LV‐siGsα^R201C‐4 and LV‐siGsα^R201C‐4.

Figure 7

Schematic representation of the interplay of Gsα and PDE activity in the regulation of in vitro adipogenic and osteogenic differentiation of BMSCs. Inhibition of PDE (IBMX) results in marked enhancement of upregulation of adipogenic markers in BMSCs induced to in vitro adipogenesis. Constitutive activity of Gsα (R201C) results in a less prominent but clear‐cut enhanced expression of adipogenic markers. Inhibition of PDE in the presence of constitutively active Gsα, in contrast, results in the block of adipogenic differentiation. In BMSCs induced to in vitro osteogenic differentiation, addition of IBMX enhances the upregulation of BSP, ALP, and RANKL. Enhanced expression of the same markers is also observed in the absence of IBMX as an effect of constitutive activity of Gsα. Magnitude of enhancement is represented schematically by red, light red, and gray character colors in the notation of the different markers.