In vitro analysis of osteogenic inhibition in non-union spinal fusion caused by nucleus pulposus cells.

Discectomy and spinal fusion is the gold standard for spinal surgery to relieve pain. However, fusion can be hindered for yet unknown reasons that lead to non-fusions with pseudo-arthrosis. Clinical observations indicate that presence of residual intervertebral disc (IVD) tissue might hinder the ossification. We hypothesize that BMP-antagonists are constantly secreted by IVD cells and potentially prevent the ossification process. Furthermore, L51P, the engineered BMP2 variant, stimulates osseo-induction of bone marrow-derived mesenchymal stem cells (MSC) by antagonizing BMP-inhibitors. Human MSCs, primary nucleus pulposus (NPC) and annulus pulposus cells (AFC) were isolated and expanded in monolayer cultures up to passage 3. IVD cells were seeded in 1.2% alginate beads (4Mio/mL) and separated by culture inserts from MSCs. MSCs were kept in 1:control medium, 2:osteogenic medium±alginate beads, 3:osteogenic medium+NPC (±L51P) and 4:osteogenic medium+AFC (±L51P) for 21 days. Relative gene expression of bone-related genes, alkaline phosphatase assay and histological staining were performed. Osteogenesis of MSCs was hindered as shown by reduced alizarin red staining in the presence of NPC. No such inhibition was observed if co-cultured with alginate only or in the presence of AFC. The results were confirmed on the RNA and protein level. Addition of L51Pto the co- cultures, however, induced mineralization of MSCs in presence of NPC. We demonstrated that NPC secrete BMP-antagonists that prevent osteogenesis of MSCs and L51P can antagonize BMP-antagonists and induce bone formation.

Investigation of bone inhibition in non-union spinal fusion caused by intervertebral disc cells in vitro

Introduction: Discectomy and spinal fusion is the gold standard for spinal surgery to relieve pain. However, fusion can be hindered for yet unknown reasons that lead to non-fusions with pseudo-arthrose. It is hence appealing to develop biomaterials that can enhance bone formation. Clinical observations indicate that presence of residual intervertebral disc (IVD) tissue might hinder the ossification. We hypothesize that BMP-antagonists are constantly secreted by IVD cells and potentially prevent the ossification process. Furthermore, L51P, the engineered BMP2 variant, stimulates osteoinduction of bone marrow-derived mesenchymal stem cells (MSC) by antagonizing BMP-inhibitors. Methods: Human MSCs, primary nucleus pulposus (NPC) and annulus pulposus cells (AFC) were isolated and expanded in monolayer cultures up to passage 3. IVD cells were seeded in 1.2% alginate beads (4Mio/mL) and separated by culture inserts from MSCs in a co-culture set-up. MSCs were kept in 1:control medium, 2:osteogenic medium+alginate control, 3:osteogenic medium+NPC (±L51P) and 4:osteogenic medium+AFC (±L51P) for 21 days. Relative gene expression of bone-related genes, Alkaline Phosphatase (ALP) assay and histological staining were performed. Results: Osteogenesis of MSCs was hindered as shown by reduced alizarin red staining in the presence of NPC. No such inhibition was observed if co-cultured with alginate only or in the presence of AFC. The results were confirmed on the RNA and protein level. Addition of L51P to the co-cultures induced mineralization of MSCs, however a reduced ALP was observed. Conclusion: We demonstrated that NPC secrete BMP-antagonists that prevent osteogenesis of MSCs and L51P can antagonize BMP-antagonists and induce bone formation.

Conditioned medium from fresh and demineralized bone enhances osteoclastogenesis in murine bone marrow cultures

OBJECTIVES Osteoclasts rapidly form on the surface of bone chips at augmentation sites. The underlying molecular mechanism, however, is unclear. Soluble factors released from bone chips in vitro have a robust impact on mesenchymal cell differentiation. Whether these soluble factors change the differentiation of hematopoietic cells into osteoclasts remains unknown. METHODS Osteoclastogenesis, the formation of tartrate-resistant acid phosphatase-positive multinucleated cells, was studied with murine bone marrow cultures exposed to RANKL and M-CSF, and conditioned medium from fresh (BCM) and demineralized bone matrix (DCM). Histochemical staining, gene and protein expression, as well as viability assays were performed. RESULTS This study shows that BCM had no impact on osteoclastogenesis. However, when BCM was heated to 85°C (BCMh), the number of tartrate-resistant acid phosphatase-positive multinucleated cells that developed in the presence of RANKL and M-CSF approximately doubled. In line with the histochemical observations, there was a trend that BCMh increased expression of osteoclast marker genes, in particular the transcription factor c-fos. The expression of c-fos was significantly reduced by the TGF-β receptor I antagonist SB431542. DCM significantly stimulated osteoclastogenesis, independent of thermal processing. CONCLUSIONS These data demonstrate that activated BCM by heat and DBM are able to stimulate osteoclastogenesis in vitro. These in vitro results support the notion that the resorption of autografts may be supported by as yet less defined paracrine mechanisms.

Recombinant Human Bone Morphogenetic Protein 9 (rhBMP9) Induced Osteoblastic Behaviour on a Collagen Membrane Compared With rhBMP2

BACKGROUND Bone morphogenetic protein 9 (BMP9) has previously been characterized as one of the most osteogenic growth factors of the BMP-family, however, up until now, these experiments have only been demonstrated using adenovirus-transfection experiments (gene therapy). With the recent development of recombinant human (rh)BMP9, the aim of the present study was to investigate its osteopromotive potential versus rhBMP2 when loaded onto a collagen membrane. METHODS ST2 stromal bone marrow cells were seeded onto 1)control; 2)rhBMP2-low(10ng/ml); 3)rhBMP2-high(100ng/ml); 4)rhBMP9-low(10ng/ml); and 5)rhBMP9-high(100ng/ml) porcine collagen membranes. Groups were then compared for cell adhesion at 8 hours, cell proliferation at 1, 3 and 5 days real-time PCR at 3 and 14 days for genes encoding Runx2, alkaline phosphatase(ALP) and bone sialoprotein(BSP) at 3 and 14 days and alizarin red staining at 14 days. RESULTS While rhBMP2 and rhBMP9 demonstrated little effects on cell attachment and proliferation, pronounced increases were observed on osteoblast differentiation. It was found that all groups significantly induced ALP mRNA levels at 3 days and BSP levels at 14 days, however rhBMP9-high demonstrated significantly higher values when compared to all other groups for ALP levels (5-fold increase at 3 days and 2-fold increase at 14 days). Alizarin red staining further revealed that both concentrations of rhBMP9 induced up to 3-fold more staining when compared to rhBMP2. CONCLUSION These results indicate that the combination of collagen membranes with rhBMP9 significantly induced significantly higher ALP mRNA expression and alizarin red staining when compared to rhBMP2. These findings suggest that rhBMP9 may be a suitable growth factor for future regenerative procedures in bone biology.

Dual Role of Mesenchymal Stem Cells Allows for Microvascularized Bone Tissue-Like Environments in PEG Hydrogels.

In vitro engineered tissues which recapitulate functional and morphological properties of bone marrow and bone tissue will be desirable to study bone regeneration under fully controlled conditions. Among the key players in the initial phase of bone regeneration are mesenchymal stem cells (MSCs) and endothelial cells (ECs) that are in close contact in many tissues. Additionally, the generation of tissue constructs for in vivo transplantations has included the use of ECs since insufficient vascularization is one of the bottlenecks in (bone) tissue engineering. Here, 3D cocultures of human bone marrow derived MSCs (hBM-MSCs) and human umbilical vein endothelial cells (HUVECs) in synthetic biomimetic poly(ethylene glycol) (PEG)-based matrices are directed toward vascularized bone mimicking tissue constructs. In this environment, bone morphogenetic protein-2 (BMP-2) or fibroblast growth factor-2 (FGF-2) promotes the formation of vascular networks. However, while osteogenic differentiation is achieved with BMP-2, the treatment with FGF-2 suppressed osteogenic differentiation. Thus, this study shows that cocultures of hBM-MSCs and HUVECs in biological inert PEG matrices can be directed toward bone and bone marrow-like 3D tissue constructs.

Extracellular Iron is a Modulator of the Differentiation of Osteoclast Lineage Cells.

Osteoclasts originate from the hematopoietic stem cell and share a differentiation pathway with the cells of the monocyte/macrophage lineages. Development and activation of osteoclasts, and as a consequence regulation of bone resorption, depend on two growth factors: macrophage colony-stimulating factor and receptor activator of NF-κB ligand. Furthermore, cell development and activity are modulated by a microenvironment composed of cytokines and growth factors and of the extracellular matrix. Membrane transporters are a means for cells to interact with their environment. Within this study, the expression of proteins regulating cellular iron homeostasis in osteoclast-like cells grown from bone marrow-derived progenitors was compared to the expression of this set of proteins by monocyte/macrophage lineage cells. In differentiating osteoclasts, levels of transcripts encoding transferrin receptor 1 and divalent metal transporter 1 (Slc11A2) were increased, while levels of transcripts encoding ferroportin (Slc40A1) and natural resistance-associated macrophage protein 1 (Slc11A1) were decreased. Supplementation of the culture media with exogenous iron led to an increase in the proliferation of osteoclast progenitor cells and to the expression of a macrophage-like phenotype, while the development of osteoclasts was reduced. Upon transfer of mature OC onto a CaP substrate, iron depletion of the medium with the Fe(3+)-chelator Deferoxamine Mesylate decreased CaP dissolution by ~30 %, which could be restored by addition of exogenous iron. During the 24 h of the assay, no effects were observed on total TRAP activity. The data demonstrate transcriptional regulation of the components of cellular iron transporters during OC development and suggests that iron homeostasis may contribute to fine-tuning of the RANKL-induced OC development.

Biology of Bone Metastases in Prostate Cancer

Advanced-stage prostate cancer (PCa) patients are often diagnosed with bone metastases. Bone metastases remain incurable and therapies are palliative. PCa cells prevalently cause osteoblastic lesions, characterized by an excess of bone formation. The prevailing concept indicates that PCa cancer cell secrete an excess of paracrine factors stimulating osteoblasts directly or indirectly, thereby leading to an excess of bone formation. The exact mechanisms by which bone formation stimulates PCa cell growth are mostly elusive. In this review, the mechanisms of PCa cancer cell osteotropism, the cancer cell-induced response within the bone marrow/bone stroma, and therapeutic stromal targets will be summarized.

AUGMENTATION OF RECONSTITUTION OF GUT ASSOCIATED LYMPHOID TISSUE IN THE SYNGENEIC MURINE BONE MARROW TRANSPLANTATION MODEL

Gut was studied as a prototypical mucosal membrane in the murine BDF-1 syngeneic bone marrow transplant model. Measures of jejunal intraepithelial lymphocytes (IELs) and crypt cells were obtained by standard techniques and a method of quantifying gut lamina propria plasma cells (PCs) was developed. The degree of ablation of gut PCs and IELs after 900 rads total body irradiation with ('60)Co, and their repopulation effected by transplantation with 2.0 x 10('5) or 1.0 x 10('6) bone marrow cells demonstrated a prolonged period of profound depression in population levels of these cells which was not reflected by the extent of damage sustained to the epithelium. Differences in the depopulation and recovery of gut PCs and IELs revealed a tendency towards initial differentiation of effector cells. A positive dose response to high bone marrow cell innocula was obtained. Subsequent studies determined that gut IEL and PC repopulation was potentiated by the addition of IELs or buffy coat cells (BCs) to the bone marrow transplant. A method of isolating 1.4 - 4.0 x 10('7) viable IELs per gram of murine small bowel was devised employing intralumenal hyaluronidase digestion of the epithelial layer and centrifugation of the resulting suspension through discontinuous Percoll gradients. Irradiated mice received 2.0 x 10('5) bone marrow cells along with an equal number of IELs or BCs. The extent and duration of depression of numbers of IELs and PCs was markedly reduced by the addition of the IEL isolate to the transplantation innocula, and to a lesser degree by the addition of BCs. The augmentation of repopuation far exceeded that expected by simple lodging of cells suggesting that the additionally transplanted cells contained a subpopulation of mucosal membrane lymphoid stem cells or helper cells. Correlation analysis of PC versus IEL levels suggests a possible feedback mechanism governing the relative size of their populations. Normal ratios of IgA, IgM, and IgG bearing PCs was maintained post transplantation with all of the regimens. ^

Tumor immunization strategies for enhancing the graft versus tumor activity of allogeneic bone marrow transplantation

Allogeneic bone marrow transplantation (BMT) is known to induce a beneficial anti-tumor immune response called graft-versus-tumor (GVT) activity. However, GVT activity is closely associated with graft-versus-host disease (GVHD), a potentially fatal immune response against antigens on normal recipient tissues. The T-cell populations mediating these two processes are often overlapping, but studies have shown that some donor T-cells can be tumor-specific. Therefore, the goal of this study was to develop strategies for preferentially activating donor T-cells capable of mediating GVT activity but not GVHD. The three hypotheses tested were: (1) Pre-transplant immunization of BMT donors with a recipient-derived tumor cell vaccine will induce a relative increase in GVT activity as compared to GVHD. (2) Post-transplant tumor immunization of BMT recipients will enhance GVT activity without exacerbating GVHD. (3) Pre-transplant immunization of BMT donors against a tumor-specific antigen will enhance GVT activity without exacerbating GVHD. ^ To test the first two hypotheses, C3H.SW mice (MHC-matched donors) were immunized with a C57BL/6 (recipient)-derived tumor cell vaccine (leukemia or fibrosarcoma) prior to BMT, or recipients were immunized starting one month after BMT. Both donor and recipient immunization led to a significant increase in GVT activity (enhanced recipient survival and decreased tumor growth). However, donor immunization also increased fatal GVHD, which was at least partially due to activation of alloreactive T-cells recognizing the immunodominant minor histocompatibility antigen B6dom1. GVT immunity following recipient immunization was not associated with an exacerbation of GVHD or a response to B6dom1. ^ To test the third hypothesis, influenza nucleoprotein (NP) was used as a model tumor antigen. C3H.SW donors were immunized against NP prior to BMT, which led to a significant increase in GVT activity. Although recipients were not completely protected against growth of antigen loss variant tumors, there was no increase in GVHD. ^ In conclusion, (1) immunization of allogeneic BMT donors with a recipient-derived tumor cell vaccine substantially increases GVT activity but also exacerbates GVHD, (2) post-transplant tumor immunization of allogeneic BMT recipients significantly increases GVT activity and survival without exacerbating GVHD, and (3) immunization of allogeneic BMT donors against a tumor-specific antigen significantly enhances GVT activity without exacerbating GVHD. ^

B cells in ADA deficiency

Deficiency of the enzyme adenosine deaminase (ADA) results in severe lymphopenia in humans. Mice with an inactivating mutation in the ADA gene also exhibit profound lymphopenia, as well as pulmonary insufficiency and ribcage abnormalities. In fact, the mouse model has a phenotype that is remarkably similar to that of the human disease, making the mice valuable tools for unraveling the mechanism of lymphocyte destruction in absence of this housekeeping gene. T cell deficiency in ADA deficiency has been extensively studied by others, revealing a block in early thymocyte development. In contrast, our studies revealed that early B cell development in the bone marrow is normal. ADA-deficient mice, however, exhibit profound defects in germinal center formation, preventing antigen-dependent B cell maturation in the spleen. ADA-deficient spleen B cells display significant defects in proliferation and activation signaling, and produce more IgM than their normal counterparts, suggesting that extrafollicular plasmablasts are overrepresented. B cells from ADA-deficient mouse spleens undergo apoptosis more readily than those from normal mouse spleens. Levels of ADA's substrates, adenosine and 2′-deoxyadenosine, are elevated in both bone marrow and spleen in ADA-deficient mice. S ′-adenosyihomoeysteine hydrolase (SAH hydrolase) activity is significantly inhibited in both locales, as well. dATP levels, though, are only elevated in spleen, where B cell development is impaired, and not in bone marrow, where B cell ontogeny is normal. This finding points to dATP as the causative agent of lymphocyte death in ADA deficiency. ADA deficiency results in inhibition of the enzyme ribonucleotide reductase, thereby depleting nucleoside pools needed for DNA repair. Another mouse model that lacks a functional gene encoding a protein involved in DNA repair and/or cell cycle checkpoint regulation, p53-binding protein 1, exhibits blocks in T and B cell development that are similar to those seen in ADA-deficient mice. Unraveling the mechanisms of lymphocyte destruction in ADA deficiency may further understanding of lymphocyte biology, facilitate better chemotherapeutic treatment for lymphoproliferative diseases, and improve gene and enzyme therapy regimens attempted for ADA deficiency. ^

Determination of the Myogenic Potential of Human Embryonic Stem Cell-Derived Mesenchymal Stem Cells

Human embryonic stem cells (hESCs) have the potential to differentiate to all adult somatic cells. This property makes hESCs a very promising area of research for the treatment of disorders in which specific cell populations need to be restored. Despite this potential, research that focuses on producing mesodermally derived cell populations from hESCs is decidedly limited, notwithstanding the prevalence of disorders involving mesodermal tissues for which treatment options are limited. Skeletal muscle myoblasts are derivatives of mesodermal cells and are characterized by the expression of the MyoD gene. These cells are difficult to obtain from hESCs in a reproducible and efficient manner. Recent developments in the field have showed some success in obtaining myogenic cells from hESCs through a mesenchymal stem cell (MSC)-like intermediate population. MSCs, which are an adult stem cell population typically derived from the bone marrow, are capable of generating multiple cell types including skeletal muscle. The aim of this study was to develop an efficient method that derives myoblasts from an MSC-like intermediate. To accomplish this goal, we first set out to isolate and expand the MSC-like intermediate from hESCs differentiated in vitro. Difficulties in reproducing published cell-differentiation methodologies, which represent a significant and familiar challenge in hESC research, are highlighted in this report.

Ontogeny and regulation of interleukin-7 expressing thymic epithelial cells

T cell development is a multistage process of differentiation that depends on proper thymocyte-thymic epithelial cell (TEC) interactions. Epithelial cells in the thymus are organized in a three-dimensional network that provides support and signals for thymocyte maturation. Concurrently, proper TEC differentiation in the adult thymus relies on thymocyte-derived signals. TECs produce interleukin-7 (IL-7), a non-redundant cytokine that promotes the survival, differentiation, and proliferation of thymocytes. We have identified IL-7 expressing TECs throughout ontogeny and in the adult thymus by in situ hybridization analysis. IL-7 expression is initiated in the thymic fated domain of the thymic primordium by embryonic day 11.5, in a Foxn1 independent pathway. Marked changes occur in the localization and regulation of IL-7 expressing TECs during development. Whereas IL-7 expressing TECs are present throughout the early thymic rudiment, the majority of IL-7 producing TECs are concentrated in the adult thymic medulla. By analyzing mouse strains that sustain blocks at different stages of thymocyte development, we show that IL-7 expression is initiated independently of hematopoietic-derived signals during thymic organogenesis. However, thymocyte-derived signals play an essential role in regulating IL-7 expression in the adult TEC compartment. Furthermore, distinct thymocyte subsets regulate the expression of IL-7 and keratin 5 in adult cortical epithelium. Intraperitoneal injection of Recombination Activating Gene deficient mice (RAG-2−/−) with anti-CD3ϵ monoclonal antibody (mAb) induces CD4− 8− double negative thymocytes to undergo β-selection and differentiate into CD4+8+ cells. Analysis of the thymic stromal compartment reveals that progression through β-selection renders thymocytes competent to alter the pattern of IL-7 expression in the cortical TEC compartment. RAG-2−/− mice do not generate mature T cells and therefore the RAG-2−/− thymus is devoid of organized medullary regions. Histological examination of RAG-2−/− thymus following anti-CD3ϵ stimulation reveals the emergence of mature thymic medullary regions, as assessed by H & E staining and expression of thymic stromal medullary markers. Stromal medullary reorganization occurs in the absence of T cell receptor αβ expression, suggesting that activation of RAG-2−/− thymocytes by CD3ϵ ligation generates thymocyte-derived signals that induce thymic epithelial reorganization, generating a mature medullary compartment. This model provides a tool to assess the mechanisms underlying thymic medullary development. ^

Molecular mechanisms of response in gastrointestinal stromal tumors treated with imatinib mesylate

Gastrointestinal stromal tumors (GIST) represent 80% of sarcoma arising from the GI tract. The inciting event in tumor progression is mutation of the kit or, rarely, platelet derived growth factor receptor-α (PDGFR) gene. These mutations encode ligand independent, constitutively active proteins: Kit or PDGFR. ^ These tumors are notoriously chemo and radio resistant. Historically, patients with advanced disease realized a median overall survival of 9 months. However, with modern management of GIST with imatinib mesylate (Novartis), a small molecule inhibitor of the Kit, PDGFR, and Abl tyrosine kinases, patients now realize a median overall survival greater than 30 months. However, almost half of patients present with surgically resectable GIST and the utility of imatinib in this context has not been prospectively studied. Also, therapeutic benefit of imatinib is variable from patient to patient and alternative targeted therapy is emerging as potential alternatives to imatinib. Thus, elucidating prognostic factors for patients with GIST in the imatinib-era is crucial to providing optimal care to each particular patient. Moreover, the exact mechanism of action of imatinib in GIST is not fully understood. Therefore, physicians find difficulty in accurately predicting which patient will benefit from imatinib, how to assess response to therapy, and the time at which to assess response. ^ I have hypothesized that imatinib is tolerable and clinically beneficial in the context of surgery, VEGF expression and kit non-exon 11 genotypes portend poor survival on imatinib therapy, and imatinib's mechanism of action is in part due to anti-vascular effects and inhibition of the Kit/SCF signaling axis of tumor-associated endothelial cells. ^ Results herein demonstrate that imatinib is safe and increases the duration of disease-free survival when combined with surgery. Radiographic and molecular (namely, apoptosis) changes occur within 3 days of imatinib initiation. I illustrate that non-exon 11 mutant genotypes and VEGF are poor prognostic factors for patients treated with imatinib. These findings may allow for patient stratification to emerging therapies rather than imatinib. I show that imatinib has anti-vascular effects via inducing tumor endothelial cell apoptosis perhaps by abrogation of the Kit/SCF signaling axis. ^

Children with acute leukemia: A comparison of outcomes and cost-effectiveness from allogeneic blood stem cell and bone marrow transplantation

The relative merits of PBSCT versus BMT for children with standard and high risk hematologic malignancies remain unclear. In a retrospective single center study, we compared allogeneic peripheral blood stem cell transplantation (PBSCT) (n=30) with bone marrow transplantation (BMT) (n=110) in children with acute leukemia. We studied recipients of HLA matched sibling stem cells, and of stem cells from alternative donors (HLA mismatched and/or unrelated) and determined whether sourcing the stem cells from PB or marrow affected engraftment, incidence of acute and chronic GvHD, and disease-free survival at 1 year. Our results show a modest reduction in time to engraftment from PB stem cells and no greater risk of GvHD, but illustrate that the severity of the underlying disease is by far the greatest determinant of 1 year survival. Patients in the BMT group had a higher treatment success rate and lower costs than the recipients of the PBSCT within the standard but not the high risk disease group, where the treatment success rate and the cumulative costs were lower in the PBSCT group compared to the BMT group. Our current incremental cost-effectiveness ratio and analysis of uncertainty suggest that allogeneic transplantation of bone marrow grafts was a more cost-effective treatment option compared to peripheral blood stem cells in patients with standard risk childhood acute leukemia disease. For high risk disease our data are less prescriptive, since the differences were more limited and the range of costs much larger. Neither option demonstrated a clear advantage from a cost-effectiveness standpoint.^

Elucidating the role of mesenchymal stem cell participation in the tumor microenvironment

To meet the requirements for rapid tumor growth, a complex array of non-neoplastic vascular, fibroblastic, and immune cells are recruited to the tumor microenvironment. Understanding the origin, composition, and mechanism(s) for recruitment of these stromal components will help identify areas for therapeutic intervention. Previous findings have suggested that ex-vivo expanded bone marrow-derived MSC home to the sites of tumor development, responding to inflammatory signals and can serve as effective drug delivery vehicles. Therefore, we first sought to fully assess conditions under which MSC migrate to and incorporate into inflammatory microenvironments and the consequences of modulated inflammation. MSC delivered to animals bearing inflammatory insults were monitored by bioluminescence imaging and displayed specific tropism and selective incorporation into all tumor and wound sites. These findings were consistent across routes of tumor establishment, MSC administration, and immunocompetence. MSC were then used as drug delivery vehicles, transporting Interferon β to sites of pancreatic tumors. This therapy was effective at inhibiting pancreatic tumor growth under homeostatic conditions, but inhibition was lost when inflammation was decreased with CDDO-Me combination treatment. Next, to examine the endogenous tumor microenvironment, a series of tissue transplant experiments were carried out in which tissues were genetically labeled and engrafted in recipients prior to tumor establishment. Tumors were then analyzed for markers of tumor associated fibroblasts (TAF): α-smooth muscle actin (α-SMA), nerve glia antigen 2 (NG2), fibroblast activation protein (FAP), and fibroblast specific protein (FSP) as well as endothelial marker CD31 and macrophage marker F4/80. We determined the majority of α-SMA+, NG2+ and CD31+ cells were non-bone marrow derived, while most FAP+, FSP+, and F4/80+ cells were recruited from the bone marrow. In accord, transplants of prospectively isolated BM MSC prior to tumor development indicated that these cells were recruited to the tumor microenvironment and co-expressed FAP and FSP. In contrast, fat transplant experiments revealed recruited fat derived cells co-expressed α-SMA, NG2, and CD31. These results indicate TAF are a heterogeneous population composed of subpopulations with distinct tissues of origin. These models have provided a platform upon which further investigation into tumor microenvironment composition and tests for candidate drugs can be performed. ^