Stem Cells for Bone Regeneration: From Cell-Based Therapies to Decellularised Engineered Extracellular Matrices

Table 1

Highlights of selected publications regarding the osteogenic potential of various cell sources.

Author, year	Model	Cell source	Treatment	Scaffold	Outcome	Criteria
Muraglia et al., 2000 [41]	In vitro	Human BMSC clones	1 week complete medium (DMEM, 10% FBS: CM) ± FGF-2 followed by 1–3 weeks in chondrogenic, adipogenic, or osteogenic medium	NA	17% (FGF−) and 34% (FGF+) displayed the potential to originate all three phenotypes induced; 80% (FGF−) and 60% (FGF+) were able to undergo osteogenesis and chondrogenesis; 3% (FGF−) and 5% (FGF+) underwent osteogenesis only; No clones underwent solely chondro- or adipogenesis	Osteogenesis: anti-osteocalcin IHC Chondrogenesis: anti-CNII IHC; Adipogenesis: Sudan Black staining
Zuk et al., 2001 [18]	In vitro	Human PLA	PLA cells at passage 1 were differentiated for 2–6 weeks in chondrogenic, adipogenic, osteogenic, or myogenic medium	NA	Chondrogenesis: positive staining for Alcian Blue and coll-II in chondrogenically differentiated cells; Adipogenesis: 42% cells were Oil red O positive.; Osteogenesis: 50% PLA cells positive for AP staining and Von Kossa staining (calcification) present in osteogenically differentiated cells, but not undifferentiated PLA cells; Myogenesis: myogenic differentiated cells expressed MyoD1 and myosin	Chondrogenesis: Alcian Blue stain and collagen II-specific mAB; Adipogenesis: Oil red O stain; Osteogenesis: AP and Von Kossa staining; Myogenesis: phase contrast microscopy, myosin- and MyoD1-specific mAB
Quarto et al., 2001 [6]	Human ()	Autologous human BM	Autologous BMSCs cultured ex vivo in CM + FGF-2	HA	All patients recovered limb function. After 27, 16, and 15 months, the patients reported no problems with the implants. Callus formation at implant site and integration with surrounding bone	Functional use of limbs. CT and radiograph to assess bone density and callus formation
Zuk et al., 2002 [19]	In vitro	Human PLA	PLA cells at passage 1 were differentiated for 9 hours (neurogenic) or 2–6 weeks in chondrogenic, adipogenic, osteogenic, myogenic, or neurogenic medium	NA	Chondrogenesis: confirmed by positive AB staining, positive IHC (KS, CS, CNIIb, and CN10); Adipogenesis: confirmed by increased GPDH, LPL, aP2 leptin, and GLUT4 activity, plus Oil red O staining; Osteogenesis: confirmed by increased matrix mineralisation and osteogenic markers versus noninduced BMSCs; Myogenesis: confirmed through increased expression of myogenic markers versus noninduced controls; Neurogenesis: uncertain. Positive detection of nestin, NSE, and NeuN. Negative for neuronal markers: GalC, GFAP, MAP-2, NF-70, ChaT, GAD65, and MBP	Chondrogenesis: IHC against KS, CNII, and CS. WB for CNII, AG, and CN10; Adipogenesis: increased Oil red O stain, GPDH, leptin, GLUT4, PPARy2, and LPL expression; Osteogenesis: matrix mineralisation, AP enzyme activity, RT-PCR (OC, CBFA-1, AP, ON, OP, BMP-2, c-fos, CNI, PTHR, RXR-a, and VDR), and WB (OP, ON, CNI, AP, RARa, and VDR); Myogenesis: RT-PCR (myod1, myf6, myf5, myosin), WB (DES, myod1, MG, MYF, and myosin heavy chain); Neurogenesis: IHC (NSE, NeuN, GFAP, and GalC), RT-PCR (nestin, ChaT, GAD65, GFAP, and MBP)
Hicok et al., 2004 [55]	Immunodeficient SCID mice	Human PLA	PLA was washed and maintained in CM followed by 3, 7, or 14 days in CM or osteogenic differentiation medium (OM)	HA-TCP	In vitro: increased AP activity in osteodifferentiated cells. Positive Alizarin Red staining in osteodifferentiated cells versus controls; In vivo: more osteoid formation in implants + PLA than implants+ noninduced BMSCs	AP activity and Alizarin Red staining (matrix mineralisation) before implantation. In vivo: H&E staining
Warnke et al., 2004 [51]	Human ()	Autologous BM	BM + BMP7 + bone mineral blocks encased in a metal cage implanted ectopically for 7 weeks	Bovine bone mineral blocks	Vital neo-bone detected at 4 weeks, after implantation. 11 days after transplantation, bone remodelling and mineralisation were detected. Jaw function (mastication) was restored by the procedure	Bone growth detected by skeletal scintigraphy following injection of radioactive tracer
Huang et al., 2005 [85]	In vitro	Matched human () AT and BM	Chondrogenesis induced by aggregate culture	NA	More cartilage-specific ECM deposited by BM cells than AT. Cells with appearance of hypertrophic chondrocytes seen in BM but not AT deposits; More than twofold greater GAG levels in BM versus AT. 500–5000x higher CN10 levels in BM versus AT deposits	Chondrogenesis: GAGs assessed by toluidine blue stain and DMMB assay, and IHC (CNII, CN10); Adipogenesis: Oil Red O staining; Osteogenesis: Von Kossa staining for calcified ECM
Im et al., 2005 [53]	In vitro	Nonmatched human AT () and BM ()	Osteogenesis induced using OM (2-3 weeks) Chondrogenesis induced through pellet/fibrin culture	NA (2D culture)	Greater AP and Von Kossa staining in BMSCs versus ADSCs. BMSCs produced more proteoglycan and CNII	Differentiation was assessed using a semiquantitative histological grading system
Kern et al., 2006 [16]	In vitro	Nonmatched human AT (), BM (), and UCB ()	Cells were cultured in OM (2.5 weeks) or adipogenic differentiation medium (AM) Chondrogenesis induced through pellet/fibrin culture	NA (2D culture)	71% BM, 79% AT, and 100% UCB samples positive for osteogenesis; 100% BM, 94% AT, and 0% UCB positive for adipogenesis; 100% samples positive for chondrogenesis	Osteogenesis: AP and Von Kossa stains; Adipogenesis: Oil red O; Chondrogenesis: Safranin O; Surface markers (CD43, CD73, CD90, CD14, CD34, CD45, CD105, CD133, CD29, HLA I, HLA II, CD106, and CD44) were also used
Sacchetti et al., 2007 [43]	Immunodeficient nih/nu/xid/bg mice	Human BM	Cultures were grown in aMEM + 20% FBS prior to implantation for 4, 7, and 8 weeks	HA-TCP + fibrin gel	BMSCs but not muscle and skin fibroblasts formed bone + BM. Human trabecular bone and periosteal cells formed bone but no BM in vivo BMSC CFU-f cells are uniquely CD146+ and can regenerate CD146+ CFU-fs in vivo	Bone and BM formation: H&E staining; CD146 (and other surface markers) assayed by FACS and tissue immunostaining
Mesimäki  et al., 2009 [7]	Human ()	Autologous AT	Cells expanded ex vivo, mixed with β-TCP in DMEM, 15% autologous serum + BMP-2 implanted ectopically for 8 months	β-TCP	After 8 months in vivo, the implant was vascularised and resembled mature bone	Bone was analysed radiologically
Evans, 2015 [17]	Rabbit femoral condylar and trochlear groove defect	Rat	Muscle and fat transduced with human BMP-2	NA	Transduced muscle implants significantly improved healing after 6 weeks	RT-PCR: OP, CNIaI, BSP, OC, AP, and CBFA1; Histology: AB and Safranin O; CT: bone remodelling
Vishnubalaji et al., 2012 [58]	In vitro	Nonmatched human AT () and BM ()	Cells were induced towards osteogenic and adipogenic fates	NA (2D culture)	Osteogenesis induced in both cell groups. Greater AP activity, mineralisation, and significantly higher levels of OC and OP in BM versus AT cells	Osteogenesis: AP, Alizarin Red S, Von Kossa stains. Calcium levels assayed; Adipogenesis: Oil red O stain; RT-PCR: aP2, PPARy, AP, AN, OP, and OC; Surface markers: CD13, CD31, CD73, CD105, CD44, CD29, CD90, CD146, CD34, CD45, CD14, and HLADR
Brocher et al., 2013 [62]	Immunodeficient SCID mice	Nonmatched human AT () and BM ()	In vitro expansion ± chondrogenic preinduction	β-TCP or no scaffold	All preinduced BM-samples generated neo-bone after 8 weeks in vivo ± β-TCP. AT samples were dependent on β-TCP for bone formation in preinduced samples producing bone in 13/18 samples	Histology: TB, Safranin O, H&E, Movat's pentachrome, and Masson's trichrome
Sándor et al., 2014 [8]	Human () craniomaxillofacial defects	Autologous AT	Expanded in vitro and embedded in a scaffold ± BMP-2	Bioactive glass or β-TCP	Successful integration with surrounding bone noted in 10/13 cases. 1 failed case due to patient nose picking	Postoperative CT and radiographs at 12–52 months follow-up
Murata et al., 2015 [20]	Porcine femoral trochlear defect ()	Autologous AT	In vitro culture of cell spheroids moulded into cylinders	NA	Results varied with regard to cartilage, but implant sites showed better remodelling of subchondral bone than control sites	Monthly CT scans; Macroscopic IRCS grading, IHC, and histology of implants at 6 and 12 months
Reinisch et al., 2015 [78]	Immunodeficient NSD mice	Nonmatched human AT (), BM (), UCB, and skin-derived cells	Cells were expended and loaded onto scaffolds. After implantation, mice were given PTH daily	HA-TCP powder	All cells had similar phenotypes in vitro, but only BM-derived cells formed bone + BM in vivoBM-derived cells had a distinct gene expression and methylation signature suggesting skeletal proclivity	Histology, H&E, and pentachrome stains; Gene expression and methylation performed using microchip arrays