Gene activated adipose tissue fragments as advanced autologous biomaterials for bone regeneration: osteogenic differentiation within the tissue and implications for clinical translation

Abstract

Cost-effective, expedited approaches for bone regeneration are urgently needed in an ageing population. Bone Morphogenetic Proteins (BMPs) stimulate osteogenesis but their efficacy is impeded by their short half-life. Delivery by genetically modified cells can overcome this problem. However, cell isolation and propagation represent significant obstacles for the translation into the clinic. Instead, complete gene activated fragments of adipose tissue hold great potential for bone repair. Here, using an in-vitro culture system, we investigated whether adenoviral transduction with human BMP-2 can promote osteogenic differentiation within adipose tissue fragments. Osteoinduction in adipose tissue fragments was evaluated by quantitative reverse transcriptase polymerase chain reaction, immunohistology and histomorphometry. BMP-2 transduced adipose tissue synthesized BMP-2 protein over 30 days peaking by day six, which significantly promoted osteogenic differentiation as indicated by increased calcium depositions, up-regulation of bone marker genes, and bone-related protein expression. Our results demonstrate that cells within adipose tissue fragments can differentiate osteogenically after BMP-2 transduction of cells on the surface of the adipose tissue. BMP-2 gene activated adipose tissue represents an advanced osteo-regenerative biomaterial that can actively contribute to osteogenesis and potentially enable the development of a novel, cost-effective, one-step surgical approach to bone repair without the need for cell isolation.

Figure 1

Adenoviral gene transduction efficiency and hBMP-2 protein expression analysis. (A) Confocal fluorescent microscopy showing the transduction efficiency of Ad.GFP in 4 mm diameter adipose tissue discs. Cells residing on the surface of the adipose tissue fragments were transduced by Ad.GFP and showed detectable positive green signals. (B) Green fluorescent signals were undetectable in non-transduced tissue discs (negative control). (C) Quantitative analysis of transduction efficiency indicated that 4.56 ± 0.51% area of transduced adipose tissue discs was transduced as GFP-positive, while 0% area of negative control tissue discs. Nine samples per group were used for the analysis. (D) hBMP-2 protein expression from media harvests every 3 days intervals until day 30. The protein synthesis level in Ad.hBMP-2 transduced tissue discs cultured in osteogenic medium (MO + Ad.hBMP-2) showed a significant increase from day 3 to day 15, and then decreased gradually until settling at a constant synthesis level from day 15 to day 30. No detectable hBMP-2 was found in non-transduced tissue discs cultured in normal medium (MN) and in osteogenic medium (MO). Nine samples per group were used for ELISA and measurements were performed in triplicate.

Figure 2

Quantitative comparisons of bone marker gene expressions between MN (normal medium), MO (osteogenic medium) and MO + Ad.hBMP-2 groups. ALP (A), RUNX-2 (B), OPN (C), OCN (D) and BSP (E) gene expressions were significantly increased in MO + Ad.hBMP-2 group when compared to MO and MN, whilst significant differences were also found between MO and MN at most time points. hBMP-2 (F) was only expressed in Ad.hBMP-2 transduced group during the culture period, of which the highest gene expression was at 2 weeks. Four tissue discs cultured in each well were harvested as one sample for RNA isolation, and 9 samples per group were used for qRT-PCR. The significance level of MO + Ad.hBMP-2 vs. MN as well as MO vs. MN, was marked on the top of column without capped line (*for p < 0.05, **for p < 0.01 and ***for p < 0.001).

Figure 3

Histological staining of untreated fresh adipose tissue discs. (A) (×2.5); (B) (×15): HE staining showing the structure of adipose tissue consisting of adipocytes (green arrow) and stromal vascular fraction (SVF) (red and blue arrows), which contains vascular tubes (red arrows) and a cellular complex including pre-adipocytes, mesenchymal stem cells, endothelial progenitor cells etc. (blue arrows). (C) Immunofluorescence of hBMP-2 revealing hBMP-2 was undetectable in untreated rat adipose tissue. (D) Alizarin red S staining showing calcium deposition was absent in untreated fresh adipose tissue. Immunohistochemistry of OCN (E), OPN (F) and Scl (G) indicating the co-localization of minimally endogenously expressed proteins of OCN, OPN and Scl in fresh adipose tissue, respectively.

Figure 4

Alizarin red S staining of adipose tissue discs cultured in MN, MO and MO + Ad.hBMP-2 groups after 1, 2 or 4 weeks of culturing. Bright red stained calcium could be clearly seen in MO + Ad.hBMP-2 and MO groups. Sections were presented as: MN: (A) 1 week, (B) 2 weeks, (C) 4 weeks; MO: (D) 1 week, (E) 2 weeks, (F) 4 weeks; and MO + Ad.hBMP-2: (G) 1 week, (H) 2 weeks, (I) 4 weeks. Scale bar = 1 mm. (J) Histomorphometric analysis of calcium depositions within the adipose tissue discs from MO + Ad.hBMP-2, MO and MN groups. Significantly higher calcium depositions were observed in MO + Ad.hBMP-2 compared to MO (**) or MN (***); calcium content of MO was significantly higher than that of MN (**); MN showed little to no calcium deposits in the tissue discs. One representative section per tissue sample was analyzed. Nine tissue samples per group/condition were used for histomorphometry.

Figure 5

Immunohistochemical staining of osteocalcin (OCN) for adipose tissue discs cultured in MN, MO and MO + Ad.hBMP-2 groups after 4 weeks of culturing. MN: (A) 1 week, (B) 2 weeks, (C) 4 weeks; MO: (D) 1 week, (E) 2 weeks, (F) 4 weeks; and MO + Ad.hBMP-2: (G) 1 week, (H) 2 weeks, (I) 4 weeks. MO + Ad.hBMP-2 group showed the most intensive staining at all the time points. The sections showing staining for osteocalcin, osteopontin (Fig. 6), and sclerostin (Fig. 7) were retrieved from the same representative specimens and the same time points from in adjacent positions in order to detect their co-expression. Scale bar = 1 mm.

Figure 6

Immunohistochemical staining of osteopontin (OPN). MN: (A) 1 week, (B) 2 weeks, (C) 4 weeks; MO: (D) 1 week, (E) 2 weeks, (F) 4 weeks; and MO + Ad.hBMP-2: (G) 1 week, (H) 2 weeks, (I) 4 weeks. The most intensive staining was presented in MO + Ad.hBMP-2 group at each time point.

Figure 7

Immunohistochemical staining of sclerostin (Scl) and morphometric analysis of the IHC staining products: OCN, OPN and Scl. MN: (A) 1 week, (B) 2 weeks, (C) 4 weeks; MO: (D) 1 week, (E) 2 weeks, (F) 4 weeks; and MO + Ad.hBMP-2: (G) 1 week, (H) 2 weeks, (I) 4 weeks. The most intensive staining was presented in MO + Ad.hBMP-2 group at each time point. (J) Morphometric analysis of the IHC staining products: OCN, OPN and Scl at four weeks. The largest staining areas of OCN, OPN and Scl were observed in the sections from MO + Ad.hBMP-2 group, followed by MO, whilst MN the smallest. One representative section per tissue sample was analyzed. Nine tissue samples per group/condition were used for morphometric analysis. The significance level of MO + Ad.hBMP-2 vs. MN as well as MO vs. MN, was marked on the top of column without capped line (*for p < 0.05, **for p < 0.01 and ***for p < 0.001).

Figure 8

Immunofluorescent staining of hBMP-2. MN: (A) 1 week, (B) 2 weeks, (C) 4 weeks; MO: (D) 1 week, (E) 2 weeks, (F) 4 weeks; and MO + Ad.hBMP-2: (G) 1 week, (H) 2 weeks, (I) 4 weeks. Immunofluorescent staining products of hBMP-2 were shown in red, and nuclei of cells were shown in blue. MO + Ad.hBMP-2 group showed strong red fluorescent for hBMP-2 signals induced by Ad.hBMP-2 transduction, whilst little to no red signals were detected in non-transduced MN and MO groups. Scale bar = 200 μm.