Lineage tracking of mesenchymal and endothelial progenitors in BMP-induced bone formation.

To better understand the relative contributions of mesenchymal and endothelial progenitor cells to rhBMP-2 induced bone formation, we examined the distribution of lineage-labeled cells in Tie2-Cre:Ai9 and αSMA-creERT2:Col2.3-GFP:Ai9 reporter mice. Established orthopedic models of ectopic bone formation in the hind limb and spine fusion were employed. Tie2-lineage cells were found extensively in the ectopic bone and spine fusion masses, but co-staining was only seen with tartrate-resistant acid phosphatase (TRAP) activity (osteoclasts) and CD31 immunohistochemistry (vascular endothelial cells), and not alkaline phosphatase (AP) activity (osteoblasts). To further confirm the lack of a functional contribution of Tie2-lineage cells to BMP-induced bone, we developed conditional knockout mice where Tie2-lineage cells are rendered null for key bone transcription factor osterix (Tie2-cre:Osx(fx/fx) mice). Conditional knockout mice showed no difference in BMP-induced bone formation compared to littermate controls. Pulse labeling of mesenchymal cells with Tamoxifen in mice undergoing spine fusion revealed that αSMA-lineage cells contributed to the osteoblastic lineage (Col2.3-GFP), but not to endothelial cells or osteoclast populations. These data indicate that the αSMA+ and Tie2+ progenitor lineages make distinct cellular contributions to bone formation, angiogenesis, and resorption/remodeling.

Inducible cell labeling and lineage tracking during fracture repair.

Mouse models incorporating inducible Cre-ERT2/LoxP recombination coupled with sensitive fluorescent reporter lines are being increasingly used to track cell lineages in vivo. In this study we use two inducible reporter strains, Ai9iCol2a1 (Ai9×Col2a1-creERT2) to track contribution of chondrogenic progenitors during bone regeneration in a closed fracture model and Ai9i UBC (Ai9×UBC-creERT2) to examine methods for inducing localized recombination. By comparing with Ai9 littermate controls as well as inducible reporter mice not dosed with tamoxifen, we revealed significant leakiness of the CreERT2 system, particularly in the bone marrow of both lines. These studies highlight the challenges associated with highly sensitive reporters that may be activated without induction in tissues where the CreERT2 fusion is expressed. Examination of the growth plate in the Ai9iCol2a1 strain showed cells of the osteochondral lineage (cell co-staining with chondrocyte and osteoblast markers) labeled with the tdTom reporter. However, no such labeling was noted in healing fractures of Ai9iCol2a1 mice. Attempts to label a single limb using intramuscular injection of 4-hydroxytamoxifen in the Ai9i UBC strain resulted in complete labeling of the entire animal, comparable to intraperitoneal injection. While a challenge to interpret, these data are nonetheless informative regarding the limitations of these inducible reporter models, and justify caution and expansive controls in future studies using such models.

Activated protein C (APC) can increase bone anabolism via a protease-activated receptor (PAR)1/2 dependent mechanism.

Activated Protein C (APC) is an anticoagulant with strong cytoprotective properties that has been shown to promote wound healing. In this study APC was investigated for its potential orthopedic application using a Bone Morphogenetic Protein 2 (rhBMP-2) induced ectopic bone formation model. Local co-administration of 10 µg rhBMP-2 with 10 µg or 25 µg APC increased bone volume at 3 weeks by 32% (N.S.) and 74% (p<0.01) compared to rhBMP-2 alone. This was associated with a significant increase in CD31+ and TRAP+ cells in tissue sections of ectopic bone, consistent with enhanced vascularity and bone turnover. The actions of APC are largely mediated by its receptors endothelial protein C receptor (EPCR) and protease-activated receptors (PARs). Cultured pre-osteoblasts and bone nodule tissue sections were shown to express PAR1/2 and EPCR. When pre-osteoblasts were treated with APC, cell viability and phosphorylation of ERK1/2, Akt, and p38 were increased. Inhibition with PAR1 and sometimes PAR2 antagonists, but not with EPCR blocking antibodies, ameliorated the effects of APC on cell viability and kinase phosphorylation. These data indicate that APC can affect osteoblast viability and signaling, and may have in vivo applications with rhBMP-2 for bone repair.

Cellular transitions and tissue engineering.

Epithelial-to-mesenchymal transition (EMT) and endothelial-to-mesenchymal transition (EndMT) describe complex changes in progenitor lineage, cell morphology, and gene expression. Stimulated by environmental cues, these cellular transitions are essential for elements of embryonic development and can be pathologically dysregulated in disease states. EMT occurs in biological processes such as gastrulation, cardiogenesis, and fibrosis. EndMT is involved in development and tissue fibrosis, but recent studies have implicated this process in musculoskeletal biology and pathology. Tissue engineering and regenerative medicine typically rely on endogenous progenitors or progenitors expanded ex vivo to repair damaged or impaired tissues or organs. The processes of EMT and EndMT may aid in elucidating new methods for reducing fibrosis and identifying novel plastic progenitor populations for tissue repair. This review will discuss the potential for EMT and EndMT to impact on tissue engineering and regenerative medicine.

Optimisation of the Factor VIII yield in mammalian cell cultures by reducing the membrane bound fraction.

In vivo, clotting Factor VIII (FVIII) circulates in plasma bound to von Willebrand factor (vWF), and the vWF:FVIII complex prevents binding of FVIII to phosphatidylserine (PS). Activation of FVIII by thrombin releases FVIII from vWF, and subsequently FVIII binds to PS exposed on activated platelets and forms the tenase complex together with clotting Factor IX. In vitro, during serum free production of recombinant FVIII (rFVIII), production cells also expose PS, and since vWF is not present to hinder interaction of secreted rFVIII with PS, rFVIII is partly associated with the cell membrane of the production cells. Recently, we showed that as much as 90% of secreted rFVIII is bound to transiently transfected production cells during serum free conditions. In this study, we investigated the effect of including vWF in the serum free medium, and demonstrate that addition of vWF results in release of active membrane bound rFVIII to the culture medium. Moreover, the attachment of rFVIII to cell membranes of un-transfected HEK293 cells was studied in the presence of compounds that competes for interactions between rFVIII and PS. Competitive assays between iodinated rFVIII (¹²5I-rFVIII) and annexin V or ortho-phospho-L-serine (OPLS) demonstrated that annexin V and OPLS were able to reduce the membrane bound fraction of rFVIII by 70% and 30%, respectively. Finally, adding OPLS to CHO cells stably expressing FVIII increased the yield by 50%. Using this new knowledge, the recovery of rFVIII could be increased considerably during serum free production of this therapeutic protein.

The B-domain of Factor VIII reduces cell membrane attachment to host cells under serum free conditions.

Factor VIII (FVIII) is an important protein in the blood coagulation cascade and dysfunction or deficiency of FVIII causes haemophilia A. Replacement therapy with exogenous recombinant FVIII (rFVIII) works as a substitute for the missing or non-functioning FVIII. The rFVIII protein has been engineered extensively throughout the years to increase the low production yields that initially were obtained from mammalian cell cultures. The scope of this work was to investigate the interaction of rFVIII with the cell membrane surface of the producing cells in serum free medium. We wondered whether binding of rFVIII to the cell membrane could be a factor diminishing the production yield. We studied the contribution of the rFVIII B-domain to membrane attachment by transfecting several constructs containing increasing lengths of the B-domain into cells under serum free conditions. We found that 90% of rFVIII is attached to the cell membrane of the producing cell when the rFVIII variant contains a short B-domain (21aa). By increasing the length of the B-domain the membrane attached fraction can be reduced to 50% of the total expressed rFVIII. Further, our studies show that the N-linked glycosylations within the B-domain have no influence on either total expression level or membrane attachment properties.