The Biological and Physical Performance of High Strength Dicalcium Phosphate Cement in Physiologically Relevant Models

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1502

The biological and physical performance of high strength dicalcium phosphate cement in physiologically relevant models

MICHAEL PUJARI-PALMER

ACTA UNIVERSITATIS UPSALIENSIS ISSN 1651-6214 ISBN 978-91-554-9885-6 UPPSALA urn:nbn:se:uu:diva-319495 2017 Dissertation presented at Uppsala University to be publicly examined in Å2001, Ångströmlaboratoriet, Lägerhyddsvägen 1, Uppsala, Friday, 2 June 2017 at 09:15 for the degree of Doctor of Philosophy. The examination will be conducted in English. Faculty examiner: Professor Christele Combes ( Inter-University Center for Research and Materials Engineering, University of Toulouse).

Abstract Pujari-Palmer, M. 2017. The biological and physical performance of high strength dicalcium phosphate cement in physiologically relevant models. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1502. 78 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-554-9885-6.

The chemical properties of calcium phosphate cements (CPCs) are very similar to the mineral phase of bone. CPCs are, consequently, very effective substrates (scaffolds) for tissue engineering; bone and stem cells attach readily, and can proliferate and differentiate to form new bone tissue. Unlike other CPCs that may remain largely unchanged in the body for years, such as hydroxyapatite, dicalcium phosphates are remodelled by the body and rapidly converted to new bone. Unfortunately, the dicalcium phosphates are also typically too weak to support load bearing in the human body. Our laboratory has recently developed a novel, high strength brushite CPC, (hsCPC), which can reach 10-50 fold higher failure strength than many commercially available CPCs. The aim of this thesis was to investigate the physical, chemical and biological performance of hsCPCs in physiologically relevant model of drug release, load bearing, osteoconductivity, and as a scaffold for bone tissue engineering. Multiple CPCs were compared in a model of screw augmentation to determine whether the physical properties of the cement, such as bulk strength and porosity, affected orthopedic screw holding strength. In an in vitro model of bone regeneration stem cells were grown on macroporous scaffolds that were fabricated from hsCPC. Drug releasing scaffolds were fabricated to examine whether the low porosity of hsCPC impeded drug release during a 4 week incubation period. The biological activity of an incorporated drug, Rebamipide, was examined after acute and chronic incubation periods. In the drug release study it was noted that the biological response to hsCPC was significantly better than tissue culture grade polystyrene, even in groups without drug. The mechanism underlying this biological response was further investigated by testing the effect of pyrophosphate, a common cement additive, on bone cell proliferation and differentiation. This thesis concludes that a high strength cement can produce significant improvement in screw augmentation strength, if there is sufficient cortical bone near the augmentation site. The hsCPC is also cytocompatible, and can support bone and stem cell proliferation and differentiation. hsCPC scaffolds stimulated osteogenic gene expression comparable to native bone scaffolds. hsCPC scaffolds are also capable of delivering drug for up to 4 weeks, in vitro. Finally, a cement additive, pyrophosphate, stimulated differentiation, but not proliferation of bone cells.

Keywords: biomaterial, bioceramic, osteoinduction, calcium phosphate, cement, osteoblast, pyrophosphate, Rebamipide, drug delivery, screw augmentation

Michael Pujari-Palmer, Department of Engineering Sciences, Applied Materials Sciences, Box 534, Uppsala University, SE-75121 Uppsala, Sweden.

ISSN 1651-6214 ISBN 978-91-554-9885-6 urn:nbn:se:uu:diva-319495 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-319495)

For my family

List of Papers

This thesis is based on the following papers, which are referred to in the text by their roman numerals.

I Pujari-Palmer, M., Pujari-Palmer, S., Engqvist, H., Karlsson Ott, M. (2015) Rebamipide delivered by brushite cement enhances osteoblast and macrophage proliferation. PLOS one, 10(5):e0128324.

II Sladkova, M., Palmer, M., Öhman, C., Alhaddad, RJ., Esmael, A., Engqvist, H., dePeppo, GM. (2016) Fabrication of macroporous cement scaffolds using PEG particles: In vitro evaluation with induced pluripotent stem cell-derived mesenchymal progenitors. Mater Sci. Eng. C Mater Biol Appl., 69:640-52.

III Pujari-Palmer, M., Pujari-Palmer, S., Liu, X., Lind, T., Melhus, H., Eng- strand T., Karlsson Ott, M., Engqvist, H.(2016) Pyrophosphate stimulates differentiation, matrix gene expression and alkaline phosphatase activity in osteoblasts. PLOS one, 11(10):e0163530.

IV Pujari-Palmer, M., Robo, C., Procter, P., Persson, C., Engqvist, H. (2017) Influence of cement compressive strength and porosity on augmentation performance in a model of orthopedic screw pull-out. Submitted to the journal the mechanical behavior of biomedical materials.

Author’s Contributions

Paper I Major part of the planning, experimental work and writing. Paper II Major part of the planning, experimental work and writing. Paper III Major part of the planning, experimental work and writing. Paper IV Major part of the planning, experimental work and writing.

Reprints were made with permission from the respective publishers.

Papers not included in the thesis

V Sladkova, M., Palmer, M., Öhman, C., Cheng, J., Al-Ansari, S., Saad, M., Engqvist, H., dePeppo, G. (2017) Engineering Human Bone Grafts with New Macroporous Calcium Phosphate Cement Scaffolds. (in prepration)

VI Pujari-Palmer, S., Lu, X., Singh, V.P., Engmanb, L., Pujari-Palmer, M., Karlsson Ott, M. (2017) Incorporation and delivery of an organoselenium antioxidant from a brushite cement. Materials letters. 197(15):115-119.

VII Pujari-Palmer, S., Pujari-Palmer, M., Karlsson Ott, M. (2016) Reduced oxidative stress in primary human cells by antioxidant released from na- noporous alumina. J Biomed Mater Res B Appl Biomater. 104(3):568-75.

VIII Tran, RT., Palmer, M., Tang, SJ., Abell TL., Yang, J. (2012) Injectable drug-eluting elastomeric polymer: a novel submucosal injection material. Gastrointest Endosc. 75(5):1092-7.

Contents

Aim of the thesis ...... 11 Introduction ...... 12 Bone Architecture ...... 15 Macroscale architecture ...... 15 Microscale architecture ...... 15 Osteoblasts ...... 15 Osteoclasts ...... 16 Osteocytes ...... 16 Nanoscale architecture ...... 16 Atomic scale architecture ...... 17 Summary ...... 17 Bone cellular biology ...... 19 Role of matrix proteins and enzymes ...... 19 Bone Matrix proteins ...... 20 Calcium and phosphate and the effect of ions ...... 23 Mineralization process ...... 25 Wound healing in bone ...... 28 Summary ...... 28 Bone Tissue Engineering and Calcium Phosphate Biomaterials ...... 30 Biomaterials ...... 30 Bone tissue engineering parameters ...... 31 Bone mechanics ...... 31 Porosity ...... 32 Calcium phosphate biomaterials ...... 35 Physical properties of hsCPC cements ...... 37 Drug delivery ...... 40 Osteoinduction ...... 41 Summary ...... 42

Results and Discussion ...... 44 Mechanical strength of hsCPC ...... 44 Porosity of hsCPC ...... 47 Correlations between compressive strength, porosity, and pull out force 47 Drug release properties of hsCPC ...... 50 Biological activity of hsCPC ...... 52 Conclusions and Future Perspectives ...... 56 Methods ...... 59 Cement/scaffold fabrication ...... 59 Chemical analysis ...... 59 Imaging ...... 60 Porosity measurement ...... 60 Mechanical testing ...... 60 Drug release ...... 61 Biological testing ...... 62 Chemiluminescence ...... 63 Svensk sammanfattning ...... 64 Acknowlegements ...... 67 References ...... 70

Abbreviations

CPC calcium phosphate cement hsCPC high strength calcium phosphate cement PBS phosphate buffered saline Col1 Collagen 1 OPN osteopontin OCN osteocalcin ALP alkaline phosphatase ACP amorphous calcium phosphate OCP octacalcium phosphate beta-TCP beta-tricalcium phosphate MCPM monocalcium phosphate monohydrate BSP bone sailoprotein PTH parathyroid hormone Ca calcium PO4 phosphate GM growth media PKC protein kinase C

Aim of the thesis

The aim of this thesis is to evaluate the performance of a new, high strength, calcium phosphate cement (hsCPC) in physiologically relevant models of drug delivery, tissue engineering, and load bearing. Unlike metal orthopedic devices, cements can be resorbed and replaced with new tissue, will not weaken nearby bone due to compliance mismatch or stress shield- ing, and can directly stimulate healing and bone tissue regeneration. Howev- er, the mechanical strength of resorbable cements is still quite poor and cements are, thus, not suitable for load bearing. A stronger cement was, therefore, developed for load bearing applications. The physical and chemical modifications that produced a higher strength CPC may have also affected the biological properties of the cement. Therefore, the performance of hsCPC was evaluated in models of cytocompatibility, tissue regeneration, drug delivery, and load bearing.

The specific aim of each manuscript is: To determine whether the hsCPC can deliver drug, and whether the cement affects the drug bioactivity (Paper I).

To evaluate the osteoconductive and osteoinductive properties of hsCPC, as a macroporous scaffold for bone tissue engineering, in vitro (Paper II).

To identify components of hsCPC that may contribute to oste- oconduction and osteoinduction and may, therefore, lead to the development of more osteoinductive iterations of hsCPC (Paper III).

To evaluate how the physical properties of hsCPC, such as compressive strength, translates into screw holding power in a model of orthopedic screw pull out (Paper IV).

11 Introduction

More than 6.3 million patients suffer traumatic injury to calcified tissues in the U.S. each year [1, 2]. Calcium phosphate cements (CPC) are effective bone void fillers, capable of bridging gaps in injured bone to facilitate healing, primarily because CPCs closely resemble the native mineral structure and composition of bone. Roughly 70% of the total dry weight of bone is composed of calcium phosphate [3, 4]. Though CPCs are biocompatible and osteoconductive (new bone tissue readily grows on/in the CPC and it can integrate into the calcified tissues), there are many areas that still require further research and development, including: a) calcium phosphates are quite weak and are, therefore, not approved for use in load bearing applications within the body, and b) many cements act simply as inert fillers that are not osteoinductive (able to actively stimulate cell recruitment and new bone tissue growth), and c) the handling properties (ease of mixing and injection or delivery, viscosity, etc.) often require optimization. A cement that ad- dresses these issues would represent a significant advance in the biomedical device field. The development process for a biomaterial, such as a CPC, is overly complex. Each parameter is interdependent, and the final product is often emergent; it is more complex than the sum of each constituent or synthesis/processing step. For example, brushite is a calcium phosphate that resorbs more quickly than hydroxyapatite, but is often mechanically weaker. The strength of brushite cement can be improved by lowering the porosity (reducing the amount of liquid), or by optimizing the reaction that produces brushite by refining the particle size of the reactants. However, as the porosity decreases and the mechanical strength increases, there is less void volume available to be occupied by drug for drug delivery. The surface area exposed to bodily fluids is also reduced and, as a consequence, the dissolution and resorption rate of the cement is expected to decrease. Thus, despite the ex- tensive knowledge that is already available on the physical and chemical properties of brushite and calcium phosphate cements, each step in the evolution of a CPC requires us to reexamine how the changes we have made to the cement formulation or processing affect the structural, chemical, and biological properties of the CPC. The primary goal of this thesis was to evaluate the performance of a high strength CPC (hsCPC) in physiologically relevant models of drug delivery, physiological load bearing ex vivo, and as a scaffold for stem cell expansion in vitro. The testing models used in paper I (drug delivery), paper II (new

12 tissues grown on scaffolds), paper III (in vitro osteogenicity screening), and paper IV (an ex-vivo model of screw augmentation failure), are intended to mimic actual conditions in vivo. Therefore, the work presented in this thesis is referred to by the term “physiologically relevant”, as opposed to absolutely theoretical, or computational, models. In paper IV the amount of force needed to cause orthopedic screw pull out was tested in synthetic cancellous bone, Sawbone©, and with bovine cortical tissue. Cements with high and low compressive strength were compared in Sawbone to determine whether an increase in bulk compressive strength translated into an improvement in screw augmentation strength. The drug release rate of hsCPC was determined in paper I, using a cyclo-oxygenase stimulating drug, Rebamipide. High strength cements typically have low porosity. There are often both advantages and disadvantages associated with low porosity. Cements with low porosity may have slower drug release rates, thereby lengthening the drug delivery period, though the amount of drug released may consequently be too dilute to be effective. Therefore, the drug release rate was measured throughout a 4-week release period and drug activity was subsequently tested in osteoblasts. With over a half million defects requiring bone graft(s) each year in the US, and the lengthy expansion time and risks associated with harvesting and growing a patient’s bone tissue in vitro/ ex vivo; mass produced, ready to use “off the shelf”, tissue engineered bone is in high demand[5]. Unfortu- nately, few studies have directly compared the performance of different material scaffolds for expanding stem cell cultures in bioreactors. There is no acknowledged “standard” synthetic material that is consistently used as a scaffold to grow new bone tissue. Calcium phosphate scaffolds are often used because osteoblasts readily attach, and new matrix proteins and mineral deposition are observed in vivo[6, 7]. However, CPC’s typically lack a critical element necessary for ex vivo and in vivo calcified tissue regeneration: macroporosity. The high strength of hsCPC allows us to incorporate macroporosity, or void spaces where blood vessels can grow and deliver nutrients, while still maintaining an optimal strength and resorption rate. In paper II hsCPC macroporous scaffolds were compared to decellularized bovine bone, with respect to cell survival, differentiation, and gene expression. Bovine bone is readily available, unlike human bone, and contains many of the same growth factors, and osteoinductive capacity. If macroporous CPC scaffolds were capable of achieving comparable osteoinduction, CPCs would have significant advantages over bone. The main advantages of CPCs including rapid and facile synthesis, the ability to mass produce patient or injury specific scaffolds instead of being limited by the availability of donor tissue, and avoiding the possibility of disease transmis- sion associated with using allograft or xenograft tissue. As a CPC is converted from an acelluar scaffold to a fully cellularized and vascularized tissue, in vivo, the CPC will undergo both dissolution and

13 resorption. CPC’s dissolve in aqueous environments, with dissolution rates that differ depending upon the type of CPC (brushite, hydroxyapatite, etc.), and the environment (pH, ionic strength, etc.). hsCPC is composed of brushite, which dissolves faster than the synthetic equivalent of bone mineral (hydroxyapatite). Osteoclast cells, whose function is to break down bone, will also actively degrade the CPC by lowering the pH (thereby increasing the solubility) and secreting degradative enzymes[7]. As the CPC dissolves, calcium and phosphate ions will be released. Ion release from CPC is sufficient to trigger multiple processes, leading to new bone growth[7, 8]. Disso- lution clearly contributes to the osteoinductive capacity of a CPC. However, little attention has been given to how additives, typically considered inert byproducts in the final cement phase, affect the osteoinductivity of a CPC. Pyrophosphate is a calcium binding agent that inhibits crystal growth and improves cement strength. As the cement dissolves and ions are released, pyrophosphate may be released. Pyrophosphate salts may also be released when osteoclasts degrade and resorb bone. In paper III we investigated the effect of pyrophosphate on osteoblasts. Pyrophosphate stimulated expression of genes involved in making new bone matrix (collagen 1, osteopontin), and mineralization (alkaline phosphate). The consequence of these changes in vivo, and the interplay between pyrophosphate and osteoinductivity will be discussed in detail later in this thesis. The first part of this thesis describes bone physiology and the requirements for regeneration and cell differentiation. In the second part the physical, biological, and chemical properties of CPCs, and osteoinduction are reviewed. The major findings are presented in the specific aims, results, and methods sections in the final part of the thesis.

14 Bone Architecture

Macroscale architecture Bone tissue contains multiple levels of organization, as shown in Figure 1. On the macro-scale two types of bone are distinguishable: very porous cancellous bone (50-90% porosity) that is filled with marrow, and much denser cortical bone (3-5% porosity)[9].

Figure 1. The architecture of bone, from macro to atomic scale. Reproduced with permission from [10].

Microscale architecture Bone is an organized aggregation of osteons. Osteons are the basic subunit of bone, and are composed of concentric rings of mineralized bone matrix surrounding a blood vessel. Three predominant cell types found in bone are also present in the osteon: Osteoblasts, osteoclasts and osteocytes.

Osteoblasts Osteoblasts produce and mineralize the protein matrix of bone. Bone matrix, or osteoid, is composed mostly of collagen. Bone formation occurs primarily via two mechanisms: intramembranous ossification or endochondral ossification. During development much of bone is first composed of cartilage and later the cartilage is converted into mineralized bone via endochondral ossi-

15 fication[11]. In intramembranous ossification osteoblasts ship calcium and phosphate carrying vesicles (matrix vesicles) to the collagen matrix surface, where the excessive levels of calcium and phosphate facilitate precipitation of new bone mineral. During fracture healing cartilage formation can be partially bypassed, depending upon the stability of the fracture, by direct mineralization of the collagen matrix by osteoblasts via intramembranous ossification[12, 13]. In all cases, the primary source of bone mineral synthesis is osteoblasts, either directly through matrix vesicle secretion, or indirectly through the secretion of matrix proteins (collagen and osteopontin), and enzymes (alkaline phosphatase, pyrophosphatase).

Osteoclasts Osteoclasts travel across bone, degrading the surface as they pass (resorption). Growth factors and matrix proteins exposed by the resorption process stimulate subsequent bone remodeling and growth[14]. The temporary bone matrix (callus) that the body uses to stabilize and bridge the edges of fractured bone, and the immature bone (woven bone) that is slowly made by osteoblasts, must both eventually be degraded so that mature adult bone (lamellar bone) can replace it. Osteoclasts are critical to bone tissue regeneration because they actively resorb immature bone that is formed during regeneration, thereby allowing mature, woven bone to be deposited. Treat- ments that inhibit or deplete osteoclasts will increase bone density because osteoblasts will continue to create new bone unabated, while the rate of bone resorption slows as the number of osteoclasts decrease.

Osteocytes Osteocytes are mature osteoblasts that have become permanently encased in bone matrix, with only a small dendritic process that extends past the matrix to a) sense the local environment including pressure sensation, ion concentration, inflammation, etc., and b) communicate with the other cell types[15]. The primary known function of osteocytes is to sense mechanical forces, particularly fluid flow shear stress, and translate these forces into signals that facilitate bone remodeling[16]. Osteocytes represent 90-95% of all cells present in bone matrix[17].

Nanoscale architecture Osteons are actually concentric rings of aligned fibers of collagen. The collagen fibers have a repetitive structure, with regularly spaced gaps where nanoscale (50 nm) hydroxyapatite crystals can form in a controlled fashion. Osteons also incorporate other matrix proteins, some of which are described

16 in the next section, that have significant roles in mineralization and remodeling.

Atomic scale architecture The proteins that comprise the matrix are themselves composed of atoms that are strung together in a specific sequence. Mutations or alterations to the sequence of a matrix protein or enzyme can significantly impact bone regeneration and mineralization. For example, mutations in the collagen 1 gene result in poorly mineralized bone and diseases such as osteogenesis imperfecta[18].

Summary Bone is mostly calcium phosphate (70%), and collagen protein (20%) Bone tissue is produced by osteoblasts and degraded by osteoclasts The presence of specific matrix proteins and enzymes are necessary for correct bone tissue development and healing

Bone cellular biology

Bone provides mechanical support during movement and produces blood cells and mesenchymal stem cells in the marrow. Bone also acts as the larg- est reservoir of calcium, storing 99% of all calcium in the body[19, 20]. When a person’s diet is calcium deficient, calcium can be mobilized from the bone. Bone is, therefore, resorbed and rebuilt based primarily upon two stimuli: to meet the body’s continual need for calcium and phosphate (chemical stimuli), as well as mechanical support during movement (mechanical stimuli).

Role of matrix proteins and enzymes All tissues are composed of cells surrounded by an extracellular matrix. The extracellular matrix is made of proteins that are a) common to most tissue types, such as collagen or elastin, and b) tissue specific proteins. The matrix proteins in bone specifically: control the rate, location and amount of mineralization serve as a template for tissue growth and regeneration serve as a direct physical connection between the hard mineral and soft tissue dissipate loading force, particularly during failure, via a self healing mechanism[21-23].

Failure in bone tissue typically occurs between mineralized collagen fibers, or between fibers and mineral, where non-collagenous matrix proteins and “self-healing” proteins are localized[22]. Therefore the matrix composition has a very significant impact on the final strength, flexibility, and regeneration potential of mature bone. There are a large number of matrix proteins, many of which are incompletely characterized. The proteins that are directly related to the work presented in this thesis are reviewed below.

19 Bone Matrix proteins Collagen 1 Collage 1 (COL1) is the most abundant matrix protein in bone. The nanoscale organization (Figure 1) is composed of triple helices that form a larger fibril structure, with repeated clefts where calcium phosphate crystals form (Figure 3). Contact with collagen 1 is sufficient to stimulate differentiation of osteoblasts and mesenchymal stem cells via integrin receptors[24, 25]. The primary role of collagen in bone is to provide elasticity and fracture toughness by absorbing force, and by connecting hydroxyapatite mineral to the bone matrix via sacrificial bonds[26-28]. Sacrificial bonds connect the bone matrix to bone mineral (hydroxyapatite), and are weaker than other tissue bonds. During loading these weaker bonds break first, and can easily reform, thereby creating a rapid turnover, easily repaired, first line of defense against catastrophic fatigue during loading. Collagen breakdown products are also chemotactic for mesenchymal stem cells, so damaged bone matrix may actively initiate regeneration[29].

Acidic proteins Acidic proteins have a large number of acidic residues, which contain phosphate (O=P-O-), hydroxyl/alcohol (C-O-), or carboxyl (O=C-O-) groups. Most acidic proteins have the same primary effect: negatively charged, oxygen containing motifs bind calcium very strongly to a) increase the local concentration of calcium and calcium/phosphate aggregates, and b) stabilize the initial crystal nucleation structure: amorphous calcium phosphate clusters (ACP). Without the stabilizing effect of acidic proteins the concentration of calcium phosphate would have to be high enough that precipitation would also occur spontaneously across the collagen matrix, instead of only occurring in specific areas in a controlled fashion, and amorphous calcium phosphate would immediately transition to less thermodynamically stable, but more kinetically favorable (faster) crystalline forms like octacalcium phosphate (OCP).

Ostepontin Ostepontin (OPN) is expressed in mineralizing tissue where it facilitates and regulates mineralization, in inflammation where it participates in the recruitment and movement of inflammatory cells, in kidney to maintain normal functions, and as both a tissue adhesive, and force dispersing molecule (via sacrificial bonds), in bone[23, 26]. In the context of mineralization, OPN specifically appears at sites of resorption, remodeling and pathological mineralization[23, 30, 31]. Animals deficient in OPN experience a 5-fold greater amount of ectopic mineralization than wild type, and this effect can be completely abolished by supplying exogenous (given as a chemical in the cell culture medium) or endogenous (produced directly by the cells, such as

20 by transfection methods that increase gene expression) OPN[31]. Further- more, OPN directly stimulates resorption of ectopic mineralization by facili- tating acidification and resorption via resident macrophage and giant cells[31]. OPN is a negatively charged matrix protein that is heavily phosphorylated on serine residues. The phosphoserine motifs bind hydroxyapatite and other calcium phosphate crystals directly, thereby preventing further growth[32]. Its calcium binding capacity is directly correlated with the amount of phosphorylation[33]. OPN can be secreted as a matrix protein, or as a soluble cytokine like molecule. Soluble OPN is chemotactic for monocyte/macrophages, which are also activated by OPN and can later become osteoclasts, epithelial, endothelial and/or mesenchymal stem cells[30, 34]. OPN contains an R-G-D sequence, which is a positively charged sequence of amino acids (arginine (R)- glycine (G)- aspartic acid (D)) that specifically bind to cell integrin receptors and facilitate chemotaxis and/or cell adhesion[35]. OPN facilitates adhesion of monocyte, macrophages and osteoclasts via integrin A5B3, and epithelial, endothelial and mesenchymal stem cells. OPN is specifically required for osteoclast adhesion during resorption, and can directly activate cytoskeletal rearrangement, membrane ruffling and resorption [36, 37] [30]. In a pol- ymerized form, in vivo, OPN is also chemotactic for neutrophils [38]. OPN is necessary for increased resorption in response to stress and hormones like parathyroid hormone (PTH), and participates in vascularization [39, 40] [41]. Knock out, or deficiencies in osteopontin reduce the fracture toughness and elasticity of bone[42]. Architecturally, OPN is localized to the cement lines in bone, where bone is actively resorbed/remodeled [30]. Antagonists that interfere with osteoclast binding to OPN can actually prevent osteoporosis and stop bone resorption[36, 37]. In fact, the amount of calcium phosphate material resorbed by osteoclasts is directly correlated with the binding strength of osteopontin to the material surface[43].

Bone Sailoprotein Bone Sailoprotein (BSP) is in the same protein family as OPN, and is a heavily phosphorylated acid matrix protein produced largely by osteoblasts. BSP functions similarly to activate and facilitate adhesion of osteoclasts and osteoblasts, resorption, but is not chemotactic for osteoclasts[44]. BSP is chemotactic for endothelial cells and can stimulate angiogenesis via activated endothelial cells[44]. Like OPN, BSP and is also found at the cement lines, and wherever bone is actively being resorbed[21, 45]. However, unlike OPN, BSP is not involved in bone resorption due to lack of load bearing. OPN knockout mice that are restricted from load bearing do not lose bone mass, while BSP knock out mice do[46].

21 Alkaline phosphatase Alkaline phosphatase (ALP/TNAP) is an enzyme that transfers phosphate between biological molecules. ALP is found in many tissues, hence it’s al- ternative name, tissue non-specific alkaline phosphatase (TNAP). In addition to bone tissue, kidney and liver also produce large amounts of ALP. Muta- tions that reduce the activity of ALP result in poorly mineralized skeleton and a deficiency in free phosphate (hypophosphatasia)[47]. ALP is present in the blood in a soluble form, on the surface of osteoblasts attached to membrane proteins (glycosylphosphatidylinositol), present in the bone matrix and attached to membrane proteins inside matrix vesicles. The main function of alkaline phosphatase is to hydrolyze phosphomonoester bonds to produce inorganic phosphate. Pyrophosphate is one such substrate that is cleaved by ALP. ALP, therefore, increases mineralization directly by increasing the concentration of phosphate, and indirectly by cleaving a strong inhibitor of mineralization, pyrophosphate[48]. ALP can also produce pyrophosphate as well, in a reverse reaction[48]. Most importantly, ALP activity (enzymatic activity) and quantity increases in response to osteogenic signals like osteogenic media, osteogenic factors (PGE2), osteogenic biomaterials, and injury/healing/remineralization signals. ALP activity is widely considered a reli- able indicator of osteogenic differentiation[47].

Pyrophosphatase Pyrophosphatase is an enzyme that cleaves inorganic pyrophosphate to produce phosphate. It is poorly characterized in the context of mineralization, and few studies have directly compared the activity and specificity of ALP to pyrophosphatase. Therefore, it is difficult to differentiate the relative contribution of each enzyme to the mineralization process. Pyrophosphate is produced normally through the metabolism of nucleotides. When adenosine triphosphate (ATP) is converted to adenosine monophosphate (AMP) a molecule of pyrophosphate is produced. As pyrophosphate builds up it is transported out of the cell by the transporter protein progressive ankylosis protein homolog, ANK. Outside the cell pyrophosphate is cleaved to produce inorganic phosphate. If ATP transport is blocked, mineralization is impaired. When ATP is supplied exogenously however, mineralization occurs regardless of whether conditions are actually osteogenic[49]. Pyrophosphatase can directly stimulate collagen expression, which in turn causes osteoblast differentiation and mineralization, though only when expressed intracellularly[50]. The underlying mechanism depends upon the intracellular ratio of pyrophosphate to phosphate. Interestingly, the effects of pyrophosphate are not inhibited by a phosphate uptake blocker (foscarnet), which suggest that even though pyrophosphate is cleaved outside the cell to produce phosphate it is not the external phosphate concentration that mediates intracellular changes that follow exposure to pyrophosphate[32, 51].

22 Calcium and phosphate and the effect of ions Bone tissue may be exposed to elevated calcium and phosphate levels from many possible sources, including injury, dissolution/resorption of bone by osteoclasts, and dissolution of implanted calcium phosphates[52]. Both ions can stimulate osteogenic gene expression, cell proliferation/differentiation, and bone tissue mineralization. While outside the scope of this thesis, other ions, including zinc (Zn), magnesium (Mg), strontium (Sr), aluminum (Al), and silicon (Si), can also stimulate osteogenic gene expression[53]. The effects of calcium and phosphate differ depending upon the cell type, assay conditions, media, and even the organic phosphate source[9, 54]. For example, stem cells exposed to elevated calcium and phosphate will still differentiate in basal media (no osteogenic factors like ascorbic acid), but will differentiate towards chondrocytes rather than osteoblasts[9]. Materials that specifically elute calcium or phosphate may mimic the inductive signals produced by bone[55, 56]. Chai et al gives concise and illuminating data (Figure 2) showing that calcium and phosphate, each administered separately, can match or exceed the effects of osteogenic media on periosteal cell proliferation and differentiation[57].

Figure 2. The effects of calcium and inorganic phosphate on periosteal cells. Calci- um (A), or inorganic phosphate (B), were added separately in growth medium and the rate of cell proliferation was determined by the amount of nucleic acid recov- ered. C2-C10 refer to the concentration of supplemented calcium (2-10uM) in growth medium (GM), P2-10 refer to the concentration of supplemented phosphate (2-10uM) in growth medium, and OM refers to osteogenic medium. Reproduced from [57].

Phosphate that is provided exogenously can directly increase osteogenic gene expression in vitro via protein kinase C (PKC), Akt (a molecular switch for osteoblast differentiation[58], RUNX2 (master regulator of osteoblast/chondrocyte differentiation and stimulator of collagen expression), Cyclin D (a switch for cell cycle and proliferation), and MAP Kinase (Erk1/2) (the signaling pathway though which almost every proliferation, differentiation, or survival signal will travel) signaling[50]. Phosphate can also directly stimulate OPN and vascular endothelial growth factor (VEGF a

23 strong inducer of angiogenesis) expression[59, 60]. Excessively high extracellular phosphate can cause osteoblast apoptosis[4]. Calcium can also stimulate osteogenic gene expression in osteoblasts, in vitro. Extracellular calcium stimulates MAP Kinase (Erk1/2) and Cyclin D1[61, 62]. Osteogenic genes that are directly stimulated by calcium include COX-2/PGE2, CBFA1, osteocalcin, osteopontin, bone morphogenetic proteins and collagen[63-65]. The rate of proliferate and/ or differentiation is also increased by extracellular calcium[4, 65]. Some of these changes occur via a calcium sensing receptor (CaR) on the cell surface[62]). Intracellular calcium increase triggers such a large number of genes, and produce such varied cellular responses, that it is easier to organize the effects of intracellular calcium changes by the originating stimuli, rather than which genes and proteins are activated. Classical stimulators of osteoblast differentiation, parathyroid hormone and vitamin D3, both act via intracellular calcium release. Uptake (endocytosis) of dentin matrix protein (DMP-1) can also trigger release of intracellular calcium and activation of MAP Kinas (p38), and RUNX2[66]. Calcium is also chemotactic for osteoblasts, stem cells and bone progenitors[4]. Physical erosion, or chemical/biochemical degradation, of calcium phosphates can produce calcium phosphate particles. These particles are often bioactive and are, collectively, referred to as basic calcium phosphate particles (BCP)[67]. BCPs are taken up (endocytosis) by multiple cell types, including connective tissue cells like fibroblasts, where they are transported through the digestive pathway to the lysosome. The acidic conditions of the endosome/lysosome cause the BCPs to dissolve and release calcium, poten- tially increasing calcium inside the cell 10-fold[68]. As a ceramic (CPC) scaffold degrades, it also dissolves to directly release calcium and phosphate ions. When the amount of calcium and phosphate exceed solubility of solution(s), they will begin to reprecipitate as amorphous, metastable (brushite, octacalcium phosphate), or stable (hydroxyapatite) precipitants. Amorphous, poorly organized precipitant is often a precursor to most mineral phases and is, therefore, often seen first, even if only for short time periods before con- verting to more stable phases. Like BCP, amorphous precipitants can also be taken up by cells and dissolve even more readily than BCPs to release calcium and phosphate. Finally, calcium pyrophosphate precipitants, present as an additive in hsCPC, can be taken up by cells and may dissolve similarly to BCP’s[69]. Calcium pyrophosphate precipitants can also assume both amorphous and crystalline forms. There are multiple crystal isoforms of calcium pyrophosphate, each with slightly different solubility, dissolution rates and biological effects[70].

24 Mineralization process The mineralization process begins with calcium and phosphate precipitating onto a collagen matrix. The levels of calcium and phosphate in the blood are typically supersaturated, and strong attraction between negatively charge phosphate and positively charged calcium ions may cause precipitation to occur spontaneously. In bone tissue spontaneous precipitation is beneficial, while precipitation in vascular tissue can lead to calcification, and heart at- tack or stroke. Of the many mechanisms in the body that prevent unintended calcification from occurring, the two most important are the presence of calcium chelators, such as inorganic pyrophosphate, and the restricted expression of mineralizing enzymes, such as alkaline phosphatase. Calcium is chelated, or bound, by inorganic pyrophosphate in the blood. Pyrophosphate deficiency causes calcification of damaged vascular tissue[71], and diseases that inactivate pyrophosphate transport result in deficiencies in, or complete failure of, skeletal mineralization[72]. Enzymes that cleave pyrophosphate to produce free calcium and phosphate, such as alkaline phosphatase, are often localized at sites of active mineralization. The localized increase in calcium and phosphate cause precipitation from supersaturated blood to form an amorphous (disorganized) precipitant. Amorphous calcium phosphate is unstable and it converts rapidly to metastable crystalline forms, such as brushite and octacalcium phosphate, before eventually forming the most thermodynamically stable form of calcium phosphate: apatite. The precipitation process is tightly controlled, and crystals form only in specific regions of the collagen fibers, in a Hodge-Petruska quarter staggered array, where gaps as small as 6-9 nm are regularly spaced[73, 74]. Clusters of positively charged amino acids are located at the C’ terminal end of collagen (Figure 3)[75]. The positive charges, clustered within 6 nm of the gap region, attract calcium and phosphate to enter and precipitate in the gap regions (Figure 3C), thereby directly facilitate crystal nucleation and growth[75]. The negatively charged acidic residues (P-O-, C-O-, O=C-O-) of collagen, and other matrix proteins such as osteopontin, bind directly to calcium to prevent precipitation outside of the gap and overlap regions, to control the size and rate of crystal growth, and can also participate directly in crystal nucleation[33, 42].

Figure 3. Organization of collagen for mineralization. Collagen fibers present positively charged residues on the C’ terminal end of the gap region (top). CryoTEM images of a single collagen fiber (bottom, A, scale bar 50nm) before (B), and after (C), soaking in supersaturated calcium phosphate solution. Discrete bands of precipitation are indicated by arrows (C). Reproduced from [73].

Before mineralization begins proteins like collagen, enzymes such as alkaline phosphatase, and ions such as calcium and phosphate, must be concen- trated at the site of mineralization. In cases of injury and fracture repair this process is accomplished via two mechanisms.

Endochondral ossification When new bone is needed the progenitors that give rise to osteoblasts, typically mesenchymal stem cells located in the bone marrow (BMSC) or perios- tium, are recruited to the site of new bone growth/repair. Early in endochondral ossification the progenitors are instead “pushed” towards chondrocyte differentiation, and cartilage production. Later in the healing process progenitors are “pushed” towards osteoblast differentiation, and osteoclasts are recruited. The osteoclasts will resorb the cartilage and osteoblasts will replace it with immature bone matrix (collagen 1) and mineralize the matrix. Tissue that is overly thick, hypoxic, or under stress, is more likely to produce cartilage and chondrocytes than forming mineralized bone and osteoblasts directly[13]. A fracture that is completely fixed, without movement or stress during healing, will not produce cartilage and will remodel directly through mineralization of collagen.

Intramembranous ossification When the volume of tissue damage is minimal, or the injury site is stabilized, replacement bone tissue can be produced without the transitory cartilage phase, by osteoblasts. Resident osteoblasts, or recruited progenitors that have differentiated into osteoblasts, will secrete and mineralize collagen 1 directly without producing cartilage. Matrix vesicles participate in this process, as well as secreted enzymes such as alkaline phosphatase.

26 Matrix vesicles (MV) Matrix vesicles (MV) are small (20-200nm diameter) pieces of the cell membrane that have pinched off to form an enclosed, protected space. MVs typically contain molecules that are critical to mineralization, such as phosphatases (ALP, pyrophosphatase, Nucleotide pyrophosphatase 1 (NPP-1)), annexins, phosphatidyl-serine (PS), matrix metaloproteases (MMPs), calcium and phosphate (pyrophosphate), and can form calcium and/or phosphate transporting channels in the vesicle membrane[76, 77]. Many of these molecules specifically require delivery via a matrix vesicle. For instance, the activity of ALP is over 500-fold higher when in contact with phosphatidyl serine and the annexins present in the matrix vesicle[78]. Matrix vesicles are required for both endochondral ossification and intramembranous ossification[47, 76]. MV specifically regulate the amount of inorganic pyrophosphate in the extracellular matrix by producing pyrophosphate from nucleotide triphosphates with enzyme ENPP-1/NPP-1, and breaking pyrophosphate into free phosphate with the enzymes alkaline phosphatase (TNAP/ALP) and pyrophosphatase (Figure 4).

Figure 4. The role of a matrix vesicle in the production of new bone mineral is shown above. Enzymes on the outer surface increase the concentration of calcium and phosphate, while proteins inside the vesicle convert the ions into a precipitant. Reproduced from [77].

When the ratio of inorganic phosphate to pyrophosphate (Pi/PPi) is high (140-70:1) hydroxyapatite will precipitate, when it is between 70:1 and 3:1 HA crystallization will be completely inhibited, and when it is low (< 2.8:1) pathological CPPD will precipitate and may cause inflammation[47, 79]. Eventually the growing mineral is released from the matrix vesicle, likely as a result of MMP mediated loss of MV integrity, the mineral is embedded into the matrix and the MMPs may continue to participate in the matrix remodeling process[47].

27 Wound healing in bone The body responds to most injuries by (Figure 5): a) forming a clot at the injury/surgery site from blood cells and platelets, b) recruiting inflammatory cells to prevent infection and resolve the wound, c) recruiting stem cells (mesenchymal stem cells from the nearby marrow, which produce cancellous bone, or osteoblast progenitors from the periosteum, which produce cortical bone and cartilage), d) formation of a soft “callus” (collagen) matrix to bridge the wound ends that is later remodeled into bone, and new blood vessels, e) secretion of fibrocartilage in the innermost region of the callus, while the outermost region undergoes direct mineralization via intramembranous ossification, f) mineralization of the cartilage by endochondral ossification to form a hard callus, g) remodeling of the hard callus into immature woven bone, h) remodeling of woven bone into mature lamellar bone[12, 15, 80].

Figure 5. Stages of bone tissue regeneration following injury. Reproduced from [12].

Summary Mineralization occurs via two mechanism o Endochondral ossification (cartilage then bone) o Intramembranous ossification (osteoid to bone) Bone is made via o Collagen secretion (osteoid) o Mineralization matrix proteins (OPN, OCN, etc.) o Deposition of matrix vesicles o Extracellular release of enzymes and ions o Localized supersaturation of calcium and phosphate o Metabolism of pyrophosphate, and activity of ALP

28 Therapeutics that improve bone healing/mineralization o Stimulate osteoblast/osteoclast/mesenchymal stem cell (MSC) proliferation o Stimulate resorption by osteoclasts o Stimulate differentiation of preosteoblasts and MSC towards bone cells o Increase matrix protein secretion o Increase enzymatic activity (ALP)

29 Bone Tissue Engineering and Calcium Phosphate Biomaterials

Tissue engineering is a research field that focuses on the creation of new tissues outside of the body (often in bioreactors). Two representative exam- ples of tissue engineering related constructs are shown in Figure 6A, a 3D printed liver phantom (left side) which is a visual representation of a liver, and Figure 6B, replacement ear tissue grown by implanting an “empty” collagen scaffold under the skin of mice and allowing native cells to populate, and cellularize the scaffold [81, 82]. Despite the ability of the human body to repair and regenerate many different tissues, the field of tissue engineering exists precisely because there are types of injuries that the body cannot repair. Bone regeneration is usually limited by: a) cellular activity (decreases with age), b) type of injury (i.e., critical defect), and c) disease states that impair regeneration (i.e., osteoporosis or diabetes).

Figure 6. Tissue engineered ear (A) and 3D printed liver phantom (B). Reproduced from [81, 82].

Biomaterials During the healing process injured tissues require a template upon which to build replacement tissue. In tissue engineering the template, or substrate, used to grow cell and tissue is called a scaffold. Scaffolds may be made from naturally derived materials, like collagen or extracellular matrix, or from synthetic materials, like polyurethane or poly lactic co-glycolic acid (PLGA). Scaffolds are part of a larger class of materials called Biomaterials. Biomaterials include, broadly, any material designed for use within the body

30 [83]. Thus, a biomaterial may be metal (e.g. cobalt chromium hip implants), ceramic (e.g. zirconia dental implants), plastic (e.g. poly tetra-floro ethylene vascular graft prosthesis), or one of many other types of materials. A successful biomaterial must not elicit an aberrant blood or inflammatory cell reaction (hemocompatible), or tissue toxicity (biocompatible). It must, however, facilitate cell attachment and survival (cytocompatible). Ma- terials that are naturally derived (organic), or are inorganic but similar to naturally produced materials (hydroxyapatite) are less likely to be recog- nized as foreign because they resemble materials already found within the body. For example, collagen and calcium phosphate bone scaffolds do not cause abnormal inflammation, coagulation or toxicity because 90-95% of bone is made of collagen (20-25% dry weight) and hydroxyapatite (70% dry weight)[3, 4].

Bone tissue engineering parameters The physical, chemical and biological properties of bone must be considered when designing an orthopedic biomaterial or implant. These properties of bone will be explained below and expanded upon in the section discussing the requirements for a CPC.

Bone mechanics Two of the most important mechanical properties of bone are the failure strength and modulus. The failure strength reflects how much stress, or force per unit area (Newtons per cm2, or megapascals (MPa)), a material can engage before failing. The modulus reflects how stiff, or resistant to defor- mation, the material is. A stiff material, like steel, will not compress or elon- gate as much as plastic or rubber, under stress. Figure 7 shows the Young’s modulus, per unit density, plotted against the failure strength, per unit density, for a comprehensive list of materials. Bone is a composite tissue that combines the rigid strength of hydroxyapatite nanocrystals and the flexibility of a collagen extracellular matrix. The modulus of cortical bone is, therefore, less stiff than pure hydroxyapatite, but more stiff than pure collagen. Inter- estingly, the failure strength of cortical bone and hydroxyapatite are roughly equivalent, while cancellous (trabecular) bone is much weaker than even pure collagen (per unit density). This is likely due to the high porosity/macroporosity found in trabecular bone.

Figure 7. Plot of the specific modulus (modulus with respect to density) and specific strength (strength with respect to density) of natural and synthetic materials. Repro- duced from [84].

Porosity Both cancellous and cortical bone have nano- and microporosity (Figure 8) that arises from the spacing between molecules, matrix fibers, cellular structures and blood vessel channels[17]. The pore size ranges from 40nm to 100um on the microscale[85]. Cancellous bone also contains macroporosity, as seen in Figure 1, which is occupied by blood and marrow and ranges between 0.5-5mm in diameter [10].

Figure 8. Porosity in bone tissue, from the atomic scale to the macroscale. Repro- duced from [17].

Porosity, defined as the percentage of volume that is not occupied by solid material (void volume), is the most critical design parameter for bone tissue

32 scaffolds [86]. Bone is a highly vascularized tissue that requires a continual, blood supply (up to 10% of cardiac output)[87]. The key factor that deter- mines the success of a scaffold is how successfully it will vascularize in vivo. Scaffolds that are thicker than approximately 1 millimeter cannot be remodeled solely by surface remodeling. Cells must enter the scaffold to completely remodel it into bone. However, diffusion can only carry nutrients 100-200um into a scaffold[88]. A vascular network must penetrate the entire scaffold, otherwise cells in the center of the scaffold will be starved of both nutrients and oxygen. Starvation and hypoxia can lead to necrosis and, eventually, infection. Macroporosity (100-500um in diameter) is required for endothelial cells to colonize the scaffold and produce blood vessels. Bone tissue scaffolds with pore size ranging from 100-400um appear to be optimal for complete mineralization and tissue integration [20, 89, 90]. Pores created by the native vascular system (haversion canals) are in the range of 100-200um. Though the 100-400um figure is widely cited, the optimal pore characteristics vary by tissue type, tissue location, animal/species, tissue loading, wound healing, and cell type. Not all pores contribute to scaffold vascularization equally. Materials may have both closed and open pores. Open pores connect with at least one, or many, other pores (Figure 9) [91]. These interconnections allow fluid and cells to permeate the entire scaffold, similar to the highly interconnected porosity of native bone. In contrast closed pores are isolated and surrounded by solid scaffold on all sides. Cells or nutrients trapped in closed pores will not be able to interact with any other part of the scaffold. Pore interconnec- tivity is just as important as the total porosity and pore size range.

Figure 9. A beta-TCP scaffold with interconnected porosity produced via a sacrificial porogen phase. Reproduced from [91].

Porosity affects the resorption and dissolution rates by increasing the available surface area in contact with liquid [92]. The diameter of the pore also has a large impact on the surface area. Smaller pores have greater surface area per unit volume. Scaffolds with smaller pores will have significantly larger surface area exposed to liquid and are, thereby, expected to resorb faster.

33 This can be seen clearly in the case of faster resorbing and dissolving beta- TCP scaffolds (Table 1) [93].

Table 1. Rate of biodegradation of calcium phosphate scaffolds with differing porosity. Reproduced from [93].

Finally, the porosity has a significant effect on the failure strength of a scaffold. As the relative volume that is occupied by air or liquid increases, the density and strength of a material decreases (Figure 10A) [94]. Just as highly porous cancellous bone is much weaker than dense cortical bone, scaffolds with greater micro and macroporosity are mechanically weaker. The size of the pores also directly affects the failure strength of a material. Beta tricalcium phosphate/ hydroxyapatite scaffolds with smaller porogen size were significantly stronger, per porosity volume, than scaffolds with larger pore sizes (Figure 10B) [95]. Porosity can be subdivided into two categories [96]. Nanoscale pores, present as air or liquid between the individual ceramic crystals (10-100nm diameter), will reduce the modulus and fracture energy without necessarily reducing the flexural (bending) strength. Larger pores (0.1-10mm) that are long, with narrow, sharp edges, dras- tically reduce the flexural strength, and are, therefore, often the source of crack initiation and propagation [96].

Figure 10. Correlation between the compressive strength of Portland cements (calcium silicates) and porosity (A). Plotted values are observed data, superimposed over values predicted by models that assume a power or exponential relation between compressive strength and porosity. (B) Compressive strength of beta-TCP/HA sintered scaffolds with different pore size. Reproduced from [94, 95].

34 Calcium phosphate biomaterials Ceramics composed of calcium and phosphate are commonly used for non- load bearing applications in vivo. The brushite and monetite cements investigated in this thesis are an example of calcium orthophosphate cements (Cax(PO4)y). A complete list of orthophosphate cements is shown in Table 2.

Table 2. Chemical properties of calcium phosphate cements. Reproduced from [97].

The physical and chemical properties of cements are determined largely by the reactants, the reaction conditions, and the crystal structure of the reaction product. On the atomic scale calcium and phosphate form a strong, rigid ionic bond. Depending upon the pH, temperature, osmolarity, solubility and saturation concentrations of the reactants and products, calcium can bond with phosphate in many different configurations. The resulting precipitants can have very different crystal structures. Figure 11 A-E shows the difference between hydroxyapatite and brushite crystals (E). In Figure 11 A-D hydroxyapatite crystals were produced with identical chemical composition and unit cell crystal structure, but with preferential crystals growth off different faces of the seed crystal to produce whiskers/needles, plates, rods, or flakes[98]. Interestingly, though hydroxyapatite is widely considered inert, recent work has revealed that altered hydroxyapatite morphology was sufficient to trigger the immune response [99]. In contrast to hydroxyapatite, brushite crystals are plate-like (Figure 11 E). The crystal size and surface area of hydroxyapatite crystals can directly affect the rate of cell differentiation and osteogenic gene expression [100, 101].

Figure 11. Morphology of chemically identical hydroxyapatite crystals (A-D) visu- alized with electron microscope. The morphology of brushite crystals used in paper I, II, and IV (E). Reproduced from [98].

Dicalcium phosphates (CaHPO4) include brushite and monetite. Brushite is formed by combining an acidic phosphate with a basic calcium source. In this thesis acidic monocalcium phosphate monohydrate (MCPM) was combined with beta-tricalcium phosphate (beta-TCP) to produce brushite, as described previously[102]. Monetite is a dehydrated form of brushite, with the same calcium to phosphate ratio, though with slightly altered crystal lattice parameters[7]. The papers in this thesis that use monetite use the same formulations as brushite hsCPC, with conversion of brushite to monetite taking place during the sterilization process (by dehydration of brushite via autoclave).

Additives During the synthesis process citric acid and disodium dihydrogen pyrophosphate are used as crystal growth inhibitors. Interestingly, both of these com- pounds are used by the body for many different biochemical reactions, including regulation of bone crystal growth.

Citric acid Citric acid is a metabolite that is abundantly produced as we turn food into chemical energy (oxidative respiration). In bone tissue citrate binds to calcium phosphate crystals, as they form, and is sandwiched between each layer of crystal[103]. Thus far there is no known uptake mechanism of exogenous citrate into cells (it is produced from metabolism directly, via the citric acid cycle), and only 1 published study can be found on the proliferation of osteoblasts response to citric acid[104].

36 Pyrophosphate Pyrophosphate is used by the body to carry and regulate the amount of free calcium in the blood and tissues. Deficits in pyrophosphate can produce pathological calcification, like arterial calcification, as a result of elevated calcium levels[71]. Pyrophosphate can increase cell proliferation, and pyrophosphate cements produce more new bone tissue than cements without pyrophosphate[105]. Finally, pyrophosphate is taken up during bone resorption by osteoclasts. Analogs of pyrophosphate that have been modified so they cannot be broken down by the body, e.g. bisphosphonates (Figure 12), are potent inhibitors of bone resorption. The exact mechanism by which these calcium binding additives control crystal growth, and improve the mechanical properties, differs based upon the type of cement and the reaction conditions (pH, temperature, concentration, etc.). Typically, the calcium binding agent will: a) bind free calcium, thereby reducing the likelihood of precipitation (crystal nucleation), b) bind to the unstable/amorphous precipitant to stabilize it and affect the crystal growth rate, or c) bind directly to crystalline nuclei to control/slow crystal growth.

Figure 12. Chemical structure of pyrophosphate, citric acid and bisphosphonate. Reproduced and modified from [106].

Physical properties of hsCPC cements Brushite and monetite cements have reported failure strengths in the range of 2-30 MPa. Optimized brushite cements can reach compressive strengths of 50-90 MPa, which are comparable to hydroxyapatite cements and stronger than human cancellous bone (Table 3)[102, 107]. The compressive strength of hsCPC ranged from 15 MPa (Paper II) for macroporous monetite scaffolds, to 35 MPa for drug delivery scaffolds (Paper I), to 56 MPa for screw augmentation (Paper IV). The reported physical properties of human bone, calcium phosphates and polymeric bone cement (PMMA) are shown in Ta- ble 3.

37 Table 3. Mechanical and chemical properties of bone, PMMA and calcium phosphate.

The compressive strength is usually determined by applying force from a single direction (uniaxial) to a fixed sample. The types of forces that a ceramic will encounter under physiological loading bearing are more complex. As a person jumps, runs, stands, or changes walking direction the skeletal system experiences compression, tension, shear, and torsional forces, in multiple directions. The hsCPC was, therefore, subjected to a more physiologically relevant model of loading; screw pull out testing. A pull out test measures the peak (failure) force applied to a screw that has been inserted into bone, before the screw is pulled from the bone. Since the metal screws are much stronger than bone, failure typically reflects physical rupture of the bone and debris surrounding the screw threads. Figure 13 shows an FEM model of screw pull out. In figure 13 the engaged trabecula are broken as the screw thread is pulled upwards (circled in red). While clinical cases of screw pull out do not arise from a single, rapid uniaxial force, the screw pull test does give an indication of the amount of energy, and force, needed to cause screw pull out. Cements that have greater holding strength are also expected to require greater force, or energy, before failing during screw pull out testing.

Figure 13. FEM model generated image of orthopedic screw threads engaging bone trabecular. Reproduced from [108].

38 Orthopedic screws are designed to fixate devices, or to support damaged bone. A patient with a fractured tibia may need to have the fragment pieces held in place so that the bone can heal correctly. An example of a fixated tibial fracture is shown in Figure 14 [109]. The orthopedic screws will act as anchors to keep a metal plate in place, which keeps the bone fragments in place.

Figure 14. Tibial fracture that has been fixated by attaching a metal plate to the bone via orthopedic screws. Reproduced from [109].

As the orthopedic screw is inserted into the bone, the screw threads may engage both cortical and cancellous bone. Since cancellous bone is much weaker than cortical bone, cancellous bone will fail under smaller loads (at lower forces) than cortical bone. This is particularly problematic in bone that lacks a cortical shell, or in poor quality (osteoporotic) bone. If loading forces exceed the holding strength of bone, the screw will be pulled out. Cement can be injected to surround the screw threads, fill the space in porous cancellous bone (Figure 15), and form a solid bridge between screw and bone[108]. In Figure 15 the trabecula are treated like empty cells in an FEM model. These “empty cells” can be filled by a cement to increase the surface area of screw that is engaged by bone/cement. Various studies have reported than calcium phosphate augmentation can improve screw augmentation strength and reduce the likelihood of pull out, with some studies claiming comparable performance to PMMA, despite CPC’s having a 5-100 fold lower failure strength than PMMA[110-112]. However, there is no published work on how the mechanical properties of the cement affect the pull out strength.

Figure 15. Model of orthopedic screw threads engaging trabecular bone and cement. The architecture of, and pores between, trabecular bone are modeled by regularly shaped cells that can be filled with cement, bone, or are empty. Reproduced from [108].

Drug delivery Regardless of whether a drug is given systemically or injected locally, there are few barriers in the body, aside from the blood brain barrier, that prevent a drug from quickly diffusing from the site of injection into other tissues. In some cases, such as when a drug is safe and effective in one organ but toxic to another organ, it is advantageous to control exactly where, how much, and how quickly, a drug is introduced to the body. Scaffolds have become an important tool for drug delivery because they can deliver drugs locally, in a controlled manner. Drug release is governed by two factors: drug dissolution and diffusion. Molecules at high concentration in a solution will diffuse to areas of lower concentration. However, the drug must dissolve before it can diffuse. Thus the dissolution step is rate limiting, and may slow drug diffusion. Mathematical models have been developed that describe the rate of release. For a non-swelling matrix, like a calcium phosphate, if the drug dissolution is faster than diffusion the Higuchi model can describe drug release, assuming only one face of the scaffold is actively eluting drug [6, 113, 114]. Taking into account the drug per unit volume (A), drug solubility (Cs), drug diffusion coefficient (D), time (t), the release rate per unit of surface area (Q) can be calculated by equation (1).

� = � � 2� − �� ∗ � Equation 1

Equation 1 can be restated more simply. If diffusion is purely fickian the percentage drug released (M/Mo) is proportional to the square root of time (time dependent), as seen in equation 2.

1/2 (M/Mo) = k * t Equation 2

40 The Higuchi model depends on diffusion, but is independent of drug concentration, and is, therefore, proportional to t1/2.The Higuchi model does not take into account geometry, or factors such as erosion/degradation of the scaffold. The Korsmeyer-Peppas model also describe drug elution from non- swelling matrices, but takes erosion and geometry into account[114]. Equa- tion 3 is used to model drug release. If the plotted exponent (n) is 0.45, 0.45 ≤ 0.89, or > 0.89, for cylinders, the mechanism of drug release is purely fickian diffusion, anomalous (combination of fickian and non-fickian), or Case II (non-fickian including relaxation, swelling, erosion, etc.), respectively. The term km is a constant related to the diffusion rate of each specific drug.

n (M/Mo) = km * t Equation 3

The adsorption rate and erosion rate directly affect drug release. Calcium phosphate is a charged matrix, therefore drugs that are also charged may bond, or be adsorbed by electrostatic attraction, to the calcium phosphate. Brushite is acidic, therefore incorporated drugs will either be positively charged or neutral inside of brushite. A strongly charged drug will likely interact via ionic bonding, while weakly charged, and even uncharged, drugs may be attracted via Van der Waals interactions. Drug delivery via brushite has been previously studied for antibiotics, growth factors, and biomacromolecules[7].

Osteoinduction Osteoinduction was originally defined as the ability of a tissue, biomolecule or compound to cause non-bone tissue to differentiate into bone[115]. Early studies stimulated ectopic bone formation by injecting crude bone extracts, containing uncharacterized biological molecules, into muscle to produce ectopic bone. The first, and most potent, osteoinductive biomolecule(s) that was identified in bone matrix was bone morphogenic protein (BMP)[116]. The concept of osteoinduction is no longer restricted to therapeutic agents, such as BMP. Calcium phosphates can directly stimulate osteoinduction[57]. A complete understanding of biomaterial stimulated osteoinduction would allow scientists to design bone scaffolds that accelerate healing, and improve integration with native bone tissue. The mechanisms underlying osteoinduction are, therefore, a very active area of research within tissue engineering. In order to be considered “osteoinductive” a material must a) recruit bone precursors- mesenchymal stem cells (MSC), b) cause differentiation of precursor cells into osteoblasts, and subsequent production of bone matrix, and c) produce ectopic bone in soft tissues[117]. While most calcium phosphates are inherently osteoconductive, due to the chemical similarity to bone mineral, the inductive capacity appears to be correlated with the dissolution rate[118]. Calcium phosphates that degrade/ dissolve faster can produce

41 more bone tissue, though faster dissolution can also impair bone tissue regeneration if the ceramic dissolves faster than bone tissue can reform[7]. The solubility and dissolution rates are listed in Table 2 for the different calcium phosphates. Calcium sulfate is considered to be among the fastest resorbing cements (Ksp 9.1*10-6), but it may not be suitable for large defects because it dissolves too quickly (faster than new bone replaces it), resulting in gaps between scaffold and bone[119]. In contrast, hydroxyapatite (HA) is very stable in vivo, with a low dissolution rate. Hydroxyapatite is chemically bioactive, i.e., it readily forms new mineral on the surface in supersaturated solutions of calcium and phosphate. However, hydroxyapatite is only able to stimulate limited biological activity. HA implanted in vivo can stimulate osteoblast differentiation and mineralization. The leach extracts from HA can also stimulate MSC differentiation, but neither HA, or leach extracts from HA, can actively recruit stem cells[120]. The biological activity of hydroxyapatite can be enhanced by combining it with a ceramic phase that dissolves much faster, like beta-TCP, to produce a biphasic ceramic. The potential mechanism of osteoinduction caused by inorganic or synthetic materials, like laboratory synthesized brushite and hydroxyapatite, involves many processes, including surface topography/ roughness/ energy, cell and protein adhesion, ion release, etc.[6, 8, 121-123]. In the context of rapidly resorbing cements like brushite, the ion release and dissolution properties are, perhaps, the most important. Ceramics that dissolve faster will produce higher levels of bioactive ions calcium and phosphate. Although calcium and phosphate ions can reprecipitate as bone mineral (chemical bioactivity), mature bone tissue requires a complex assortment of matrix proteins produced by osteoblasts, and organized very differently than mineralization resulting from reprecipitation. Therefore, it is the biological effects of ion release on osteoblasts and stem cells, not the chemical effect on mineral precipitation, that determine the osteoinductivity of a material[57]. This also explains why the dissolution rate appears to correlate with the tissue forming potential of a cement[6, 63, 124, 125].

Summary Tissue engineering uses biomaterials to grow tissues and organs outside the body A biomaterial must be cytocompatible (non-toxic) and permissive for cell attachment, proliferation and differentiation The physical properties of the scaffold strongly affect performance in the body. o Highly porous scaffolds are needed for angiogenesis and blood supply, though the mechanical strength decreases as porosity increases

42 Each of the physical properties of a biomaterial are inter-related. o An increase in porosity will also increase the surface area, thereby increasing the drug release and resorption rate, while decreasing the mechanical strength. Scaffolds are biomaterial constructs that serve as a substrate for cell attachment, and template for the final tissue product (e.g. collagen scaffold for ear tissue)

43 Results and Discussion

Mechanical strength of hsCPC The mechanical strength of a CPC depends upon a large number of factors. In paper IV the compressive strength was varied by changing the liquid to powder ratio, using 500uM citric acid as the liquid, and the ratio of beta-TCP to MCPM (25%, 55% or 75% beta-TCP) (Figure 16). The corresponding changes in porosity (Figure 19B) were inversely related to strength, with the exception of the 8 MPa cement (75% beta-TCP, 0.3 L/P).

Figure 16. Compressive strength (A) and porosity (B) of different formulations of brushite, including hsCPC. Reproduced from paper IV.

The performance of brushite in a physiologically relevant testing regimen, screw pull out, depended upon the surrounding bone quality, as well as the cement strength. The strength of the cement had little effect on screw pull out (Figure 20A) if the surrounding bone was weaker than the cement/screw bond. The cements are significantly stronger than cancellous bone (2-56 MPa for cement, versus 0.5-10 MPa compressive strength for cancellous human bone), therefore, synthetic cancellous bone (Sawbone) failed, before the cement did. This phenomenon occurred in both simulated “healthy” (0.24 g cm-3) and “osteoporotic” (0.12 g cm-3) Sawbone. Nearby cortical bone can enhance the strength of cancellous bone during pull out testing. When cortical bone is in close proximity to cancellous bone they act as a composite material, with the aggregate failure strength determined by which layer is thicker/denser[126]. The contribution of cortical tissue to pull out strength was investigated by making artificial “cortical shells” that were much stronger or much weaker than cortical bone (Figure 17C). Fresh cortical tissue from bovine tibia was also included for comparison. The approximate failure strength of each material was 500N (polymer shell), 1000-1500N (cortical bone shell), and >>2000N (steel shell). The

44 polymer shell provided little benefit, with only a modest increase in the maximum pull out force (Figure 17B). The steel shell (Figure 17C) significantly increased the maximum pull out force, with stronger cements reaching 1000- 1500N, while weaker cements produced pull out forces of 500-700N. When cortical bone shells (Figure 17E) were placed between the Sawbone and steel testing rig the pull out forces were actually higher, for all samples, than with steel shells (Figure 17D). This reinforcement effect was also seen with polymer shell samples, where the pull out forces for all cement groups increased because the shell increased the density/resistance to pull out for all samples. Cortical bone fixated samples produced up to 1900N of force (average 1338N) with the strongest cement, though the performance of the cement was limited by the strength of the cortical bone tissue (1000-1500N). The pull out strength can be increased by increasing the volume, as shown in Figure 17C and 17D. The augmentation volume used in paper IV was quite small (1 mL) and, therefore, far higher pull out strength may be reached with larger volumes, in vivo. The strongest formulation (56 MPa) exhibited filter pressing, therefore, it is likely that the strength can be improved even farther by optimizing the injectability, viscosity and spreading distance into porous cancellous bone. The pull out forces observed in paper IV approach or match the pull out forces reported in human cadaveric bone (800-3000N)[126].

Figure 17. Comparison of the pull out strength of synthetic bone augmented with different strength cements, in healthy (A) or osteoporotic (B) Sawbone without cortical fixation. The pull out strength increased significantly in healthy Sawbone with physiologically weaker (polymer, C), equivalent (bovine cortex, D), or stronger cortical shells (steel, E). Reproduced from paper IV.

45 In paper I the cement formulation, and liquid to powder ratio, of the cement roughly corresponding to the 56 MPa group from paper IV. The mixing/ handling procedures were slightly different between paper I and IV, which produced lower compressive strengths (35 MPa) in paper I (Figure 18). Spe- cifically, the cement was molded into cylinders in paper I using a spatula to mix and shape the cement paste, while in paper IV the cement was packed into a syringe and injected through a 13 gauge needle to simulate cement delivery in a clinical setting. The incorporation of Rebamipide also reduced the strength of the cement, with a significant reduction occurring between 2 - 4% (weight percentage) (20 MPa). The XRD spectra and phase composition was comparable for all samples, which suggests that the reduction in strength occurs in the absence of a phase change, perhaps occurring via macro-structural disruption of the crystal structure, caused by the drug.

Figure 18. Compressive strength and phase composition of drug loaded hsCPC. hsCPCs were loaded with 0.5% - 4% Rebamipide and the compressive strength was determined after setting for 24 hours, and after leaching for 4 weeks (A). The phase composition of each group was determined by XRD and Rietveld refinement after 24 hours and 4 weeks (B). Reproduced from [127].

In paper II the porosity was intentionally increased, which produced significantly weaker scaffolds/cements. While the formulation (beta-TCP to MCPM ratio) of the cement remained the same between paper I, II, and IV, the amount of liquid was increased, from 0.22 to 0.25, in paper II. Consider- ing the multiple changes that each reduce the compressive strength (phase change from brushite to monetite, incorporation of macroporosity, increase in liquid to powder ratio), the macroporous monetite scaffolds still demon- strate significant strength, within the range of human cancellous bone (0.5- 10 MPa).

46 Porosity of hsCPC In paper II the porosity of hsCPC was varied to produce a scaffold that sup- ports human stem cell attachment. Macroporosity was created with a sacrificial porogen, poly-ethylene glycol (PEG). As the total porosity, and macroporosity increased the compressive strength decreased (Figure 19A, B). The optimal amount of scaffold porosity is believed to be around 60-80% total porosity, which ensures complete perfusion of the scaffold with nutrients. The scaffolds produced in paper II contained 60-75% total porosity, of which 30-60% was macroporosity (Figure 19C, D).

Figure 19. Scaffolds with macroporosity for tissue engineering. Increasing amounts of porogen increased the amount of macroporosity (A) and decreased the compressive strength (B) of hsCPC. The total porosity (C) and percentage of total porosity that was macroporous were correlated with final compressive strength. Reproduced from [128].

Correlations between compressive strength, porosity, and pull out force In paper II and paper IV the relationship between the porosity and compressive strength was investigated further. In paper IV the porosity was plotted as “inverse porosity”, or rather as the inverse value. As the porosity increases the “inverse porosity” value decreases. The correlation between these two factors was relatively weak for total porosity (Figure 19E) and only moder- ately strong for macroporosity (Figure 19D), in monetite scaffolds (paper II)[128]. This is likely because a) the scaffold structural integrity was impaired by the phase conversion from brushite to monetite, and b) large pores

47 have a larger impact on crack initiation such that macroporosity, but not total porosity, remained a strong indicator of failure. Larger, and narrower, pore morphology greatly increases the likelihood of crack initiation and propagation as well. Paper II revealed that macroporosity above 40%, and/or total porosity above 65% will produce scaffolds far weaker than surrounding cancellous bone (< 1 MPa), in vivo. In paper IV the compressive strength and total porosity were compared in brushite hsCPC. The correlation between pull out and compressive strength (Figure 20A), or inverse porosity (Figure 20B), was determined in the absence of Sawbone by testing the strength of the screw/cement bond (Figure 20C). The correlation between pull out and compressive strength (Figure 20D), or inverse porosity (Figure 20E), was determined in “healthy” (0.24 g cm-3), cortically fixated Sawbone. Interest- ingly, as the strength of cortical fixation increased the linear correlation between inverse porosity and pull out became stronger (R2 = 0.56, 0.98, 1.00 for polymer, cortical tissue, and steel shell fixation, respectively). The correlation between compressive strength and pull out, for both Sawbone fixated and screw/cement testing, was non-linear, with stronger correlation strength in Sawbone, compared to screw/cement samples. Unlike porosity, compressive strength is a complex variable, which is, itself, determined by other factors including porosity, cement viscosity and spreading inside of Saw- bone, variability between Sawbone samples, etc.

Figure 20. Correlations between the pull out force and compressive strength, or porosity without (A, B) or with Sawbone (D, E). A simple model that examined how cement strength (A) or porosity (B, represented by inverse porosity) affected the screw bonding strength was conducted without Sawbone. The sample and testing procedure is shown in C. The correlation of pull out force with compressive force (D) or total porosity (E) for Sawbone with cortical fixation. Reproduced from paper IV.

The major finding of paper IV was that porosity and compressive strength were very strong indicators of pull out force (R2 = 0.8-1.0), for calcium phosphate cement, as long as the material surrounding the cement (e.g. cancellous bone, metal plates, cortical bone) was significantly stronger than the cement. While stronger cements do produce greater pull out force, porosity is a slightly more effective predictor of performance. Within the same class of materials (dicalcium phosphates, under wet conditions), a clinician may eventually be able predict the relative performance of other cements compared to their current favored materials, simply by comparing the porosity or compressive strength to the model proposed in paper IV. In the future, after the correlations presented in paper IV are validated and refined further, a clinician may be able to select cements specific to the needs of each patient, given information such as the bone mineral density and cortical thickness of a patient. The results of paper IV were limited to a single type of calcium phosphate (brushite), and samples that deviated from the typical composition by retaining excess unreacted MCPM or beta-TCP (the 8MPa cement group) also deviated from the trend by exhibiting both increased porosity and increased strength. Therefore, future work is needed to extend to current findings to calcium phosphates with different phase compositions.

49 It should be noted that the bone model used, Sawbone, also does not completely reproduce the physiology of actual bone. Sawbone is a polyurethane foam, with open cells that allow the cement to spread through the material as it would in real bone, with compressive and shear strengths similar to real bone (according to the manufacturer). However, Sawbone differs from actual bone in many ways: the trabecular thickness and strength differ from real bone; the foam is empty and does not impede cement penetration while in actual bone the open cells are occupied by marrow, which greatly impede cement spreading. It has also been suggested that the engagement of screw threads by trabecula may transfer load throughout the trabecula network, which also may not occur in the Sawbone. Despite these limitations Sawbone remains a commonly used bone model in screw pull out testing.

Drug release properties of hsCPC Rebamipide (Figure 21A) is an anti-ulcer drug that can stimulate prostaglan- din production in select cell types[129]. In addition to scavenging hydroxyl radicals directly, Rebampide can reduce inflammation and degeneration in cartilage, and cartilage cells (chondrocytes), inhibit osteoclast differentiation, and stimulate proliferation and regeneration in mucosal cells[130, 131]. The exact signaling mechanism, or receptor, that is activated by extracellular Rebamipide is not known. Rebamipide is commonly used as an anti-ulcer, muco-protective agent[132]. In paper I we investigated whether Rebamipide could also stimulate osteogenic differentiation or proliferation in osteoblasts. Before examining the biological activity we first characterized the drug release rate. If drug was released too slowly, or incompletely because it was adsorbed to the cement or trapped inside closed pores, then hsCPC would not be a suitable carrier for Rebamipide. The initial release was linear, and predictable (Figure 21B), and release continued for up to 4 weeks. The most effective concentration, in vitro, of Rebamipide was 400uM for radical scavenging, and 0.4 and 75-100uM for osteoblast proliferation and differentiation. For each milliliter of release liquid (i.e. blood, water, etc.) approximately 37 ug of Rebamipide need to be released to stimulate osteoblast proliferation and differentiation. The concentration of released Rebamipide was roughly 19, 38, 76, and 150 ug per hour for the 0.5%, 1%, 2% and 4% loaded samples, respectively, as shown in Table 4. Therefore, hsCPC can continually deliver an effective dose of Rebamipide throughout the first 5 days.

Figure 21. The chemical structure (A) and drug release rate of Rebamipide (B) from hsCPC. Reproduced from[127].

Table 4. Drug release rate and modeling of Rebamipide release from brushite. Re- produced from [127].

The major finding of Paper I was that a) drug release from hsCPC reached biologically effective concentrations throughout the first 5 days (8-10% of loaded amount per day), which was confirmed via a radical scavenging assay, and b) even after lengthy release times (10-30 days), Rebamipide stimulated biological activity (proliferation/differentiation) in osteoblasts. The drug release profile included two distinct phases: a rapid and linear release rate during the first 5 days (Figure 21B), and slower but still linear release rate over the following 3 weeks. In retrospect, there were additional factors in Paper I that could have been investigated to produce a more accurate model of drug release. An alterna- tive form (equation 4) of the Higuchi equation, taking into account tortuosity (τ), porosity (ε), diffusion coefficient (D), amount of drug per unit volume (A), drug solubility (Cs), and time (t), may have given a better model of drug release[113].

51 !! Q = ( )(2A − ε� )(� t) Equation 4 ! ! !

The porosity of the cements should also have been characterized, partially to account for physical changes to the scaffold resulting from drug incorporation, and partially as a necessary variable to better model drug release and degradation. Finally, the degradation rate should have also been investigated as brushite is a rapidly degrading material, and because the micro and micro- scopic flaking, dissolution, erosion and general deterioration of the scaffold will increase the available surface area for diffusion/ drug dissolution. It is likely that the latter phase of Rebamipide release included contributions from brushite dissolution/degradation.

Biological activity of hsCPC hsCPC scaffolds supported human stem cell proliferation and differentiation, similar to the controls (bovine bone scaffolds) (Figure 22A), in paper II. Osteogenic gene expression was also comparable between cells seeded on hsCPC scaffolds and control bovine bone scaffolds (Figure 22B). Collective- ly, the results of paper II suggest that despite physiochemical changes arising from the unique formulation that produces hsCPC, the cement remains osteoconductive. The major finding of Paper II was that synthetic scaffolds, which can be mass produced at low cost, can support human stem cell proliferation and differentiation comparable to decellularized bone.

Figure 22. Morphology and viability (A), and long term gene expression (B) of human stem cells. Reproduced from [128].

When hsCPC was combined with Rebamipide the leach extracts stimulated proliferation in osteoblasts (Figure 23A)[127]. The cement extracts were mitogenic (group 0), and hsCPC with the lowest concentration of Re- bamipide (0.5%, weight percentage) stimulated proliferation by 50-100%. The major biological finding of paper I was that a) Rebamipide affected osteoblast proliferation and differentiation (osteogenic), b) leach from hsCPC was osteogenic, and c) the osteogenic effects of hsCPC and Re- bamipide were additive, as seen in Figure 23A. hsCPC contains pyrophosphate, an ion that participates in bone mineralization in the body. Since the leach extracts from cement contain many other ions and precipitants we looked specifically at the biological effects of pyrophosphate in paper III. Pyrophosphate directly stimulated alkaline phosphatase activity (Figure 23B, paper III), and enhanced osteogenic gene expression in osteoblasts.

Figure 23. Proliferation (A) and ALP activity (B) of MC3T3 preosteoblasts after exposure to leach products (A), or pyrophosphate (B). Reproduced from [69, 127].

This finding was not entirely novel, as the osteogenic effects of pyrophosphate have recently been studied in osteoblasts[32]. However, in paper III we demonstrated that the gene expression of important bone matrix proteins, Collagen I, Osteocalcin and Osteopontin, were upregulated by pyrophosphate exposure (Figure 24). It is possible that pyrophosphate stimulates alkaline phosphatase, and other phosphatases, by a feedback loop. When we expose cells to high levels of pyrophosphate the osteoblasts likely interpret the elevated levels as a deficiency in the enzyme that normally break down pyrophosphate. The cells respond by increasing production and gene expression of these enzymes, and that could be why alkaline phosphatase expression increases. If correct, this interpretation suggests that pyrophosphate is not osteogenic, rather it simply triggers a feedback loop by osteoblasts. This explanation is particularly attractive because all other in vitro tests have shown that pyrophosphate exposure (>1uM) actually prevents mineralization very effectively, regardless of other osteogenic effects. Paradoxically, nu- merous in vivo studies have shown that fast resorbing calcium phosphate cements (monetite) that contain pyrophosphate produce more bone and bone matrix (Collagen 1), and appear to resorb slower than cements without pyrophosphate[105, 133, 134]. This paradox could be explained by many different possibilities: a) pyrophosphate could directly produce long term changes in bone matrix, which may change how the body coats and responds to the calcium phosphate surface, thus pyrophosphate alone (delivered as an exogenous drug or precipitant) acts differently, in vitro, from pyrophosphate in a cement, in vivo; b) pyrophosphate may also act indirectly by changing the physical properties of the cement (reduce resorption rate, decrease crystal size or increase crystal packing density, changing the local solvation environment during dissolution), thereby changing the rate and amount of osteoinductive ions (calcium, phosphate) that are released; c) the rapidly dissolving monetite and brushite may dissolve, leaving behind a layer of undis- solved calcium pyrophosphate, which would be the most stable surface available for cells to interact with, therefore the in vivo properties might reflect the cell response to a calcium pyrophosphate surface. There are many

54 other possible ways to explain the contradicting in-vitro and in-vivo results. However, without additional data it is difficult to determine which explanation(s) are accurate.

Figure 24. Gene expression in mouse calavarial pre-osteoblast, MC3T3, following exposure to pyrophosphate. Reproduced from[127].

55 Conclusions and Future Perspectives

The focus of this thesis was to characterize the performance of hsCPC in physiologically relevant models. The hsCPC proved to be an effective drug delivery vehicle, for up to four weeks, in paper I. hsCPC scaffolds were effective substrates for human stem cell expansion in paper II. The leach extracts were found to be osteogenic in paper I, and an additive, pyrophosphate, was found to stimulate osteogenic gene expression in paper III. Final- ly hsCPC was able to support physiological loads in a model of screw pull out in paper IV. The hsCPC formulation investigated in this thesis appears to be an effective vehicle for multiple clinically relevant applications. This thesis makes the following specific contributions to the current scientific knowledge on CPCs: a) A significant improvement in screw augmentation strength can be obtained by optimizing the physical properties of a CPC; b) low porosity CPCs, such as hsCPC, can still effectively deliver drug over extended time periods; c) the drug, Rebamipide, and inorganic additive, pyrophosphate, both enhance alkaline phosphatase production in osteoblasts; d) preliminary evidence presented in paper III suggest that pyrophosphate may contribute to the osteoconductive, and osteoinductive, potential of a CPC; e) the altered physical properties (e.g. low porosity) of hsCPC do not impair cell adhesion, differentiation or survival. The future perspectives for calcium phosphates remain very bright. Of the hundreds of biomaterials investigated each year, many of which are “novel”, the majority of publications report the material to be cytocompatible, non- toxic, and usually suited for a particular clinical or scientific application. The number of materials that actually continue onto use in humans, thought, is remarkably small. In the case of calcium phosphate the trend is quite different. There are currently hundreds of calcium phosphate bone void fillers available, though very few actually show equivalency to autologous bone according to a recent study[135]. Since calcium phosphates integrate so easily into calcified tissues, compared to other materials, a very real challenge facing the field is to move beyond incremental improvements and broadly appropriate clinical applications, such as void filling. The biology, and biological response to cement, remain poorly understood. Many biomaterials are reported to be “osteogenic” based upon increased gene expression of collagen, alkaline phosphatase and osteopontin. However, synthetic materials like hydrogels and polymers can stimulate matrix secretion (collagen, osteopontin) because the body identifies the surface as foreign, tries to make

56 the surface more cytocompatible (collagen), and initiates a resorption response to replace the surface (osteopontin) with more biologically appropriate matrix proteins. Likewise, pyrophosphate appears to be osteogenic because it stimulates matrix proteins and genes involved in mineralization. However, these same processes are also activate during pathological calcification as well. It is unknown whether osteoblasts interpret pyrophosphate precipitant as an aberrant surface that requires osteopontin deposition to trigger remodeling. The paradox of why pyrophosphate precipitant trigger many of the genes for mineralization to occur, while also simultaneously preventing mineralization from occurring in vivo also remains unresolved. The actual biological mechanisms and effects related to pyrophosphate delivery, uptake, and signaling are poorly understood. It is not known how pyrophosphate enters the cell, how it changes the cement properties and dissolution rate, whether it triggers cell attachment receptors, such as integrins, and whether the proteins that are produced in response to pyrophosphate act chemotactically to recruit other cell types, which subsequently affect resorption. The participation of multiple cell types has not been investigated in vitro. The participation of, and effect of pyrophosphate on, osteoclasts is completely unknown. In vivo, multiple cell types, such as fibroblasts, osteoclasts, monocytes, macrophages, and stem cells, may all participate in the response to pyrophosphate. In vitro testing cannot resolve the simultaneous interactions between all these cell types. Finally, further testing is required to confirm the actual performance of hsCPC in vivo, and under physiological loading in vivo. While screw pull out is an unsolved clinical challenge, Sawbone testing may not adequately predict in vivo performance. This is particularly important given the low variability between Sawbone samples and the high variability seen between patients, and even between bone taken from different locations within the same patient. Also, simply increasing the bulk strength of the cement may not produce a measureable difference in vivo. Calcium phosphates, like most ceramics, are brittle materials and failure often occurs rapidly following crack initiation and propagation. The most common academic and industrial response to this problem has been fiber reinforcement. When fibers are mixed into the cement they can interrupt crack propagation and accept ten- sile and bending load transfer, which typically exacerbate cracks more than compressive loading. Our laboratory has recently shown that the higher strength hsCPC can last up to 5 million cycles of loading, which included flexion and bending forces, at 13 MPa of pressure[136]. The loading used by Ajaxon, et al., is quite close to the average forces seen in the hip during routine movement and load bearing[137]. Thus, with a strong enough cement and small enough routine loading force/pressure, hsCPC may actually support load bearing long enough to facilitate healing in vivo (millions of loading cycles). If hsCPC can also continually deliver anabolic bone stimulating agents (drug delivery), support stem cell attachment, and remodel rapidly

57 enough to be converted into bone at, hsCPC may represent a significant ad- vancement in the development of calcium phosphate cements.

58 Methods

Cement/scaffold fabrication Brushite hsCPC’s were formulated by combining beta-TCP and MCPM at molar ratios of 55:45. Citric acid was added as the liquid at 0.5M, at a liquid to powder ratio of 0.22 (paper I), 0.25 (paper II), or 0.22-0.3 (paper IV). Disodium dihydrogen pyrophosphate was added to the powder as 1% of the beta-TCP and 1% of the MCPM powder (weight percentage). The beta-TCP already contained approximately 10% (weight percentage) calcium pyrophosphate as an impurity. For compression and drug release testing cements were cured in cylindrical silicone molds, 6 mm diameter by 12 mm long, in 100% humidity for 24 hours. Macroporous scaffolds (paper II) were fabricated by mixing polyethylene glycol (PEG) particles (100-600um in diameter) into the cement powder, and subsequently leaching the PEG for 24-72 hours. Monetite scaffolds were created by wet sterilizing brushite at 150oC, which causes a phase change from brushite to monetite.

Chemical analysis The phase composition of hsCPC (paper I, II) and pyrophosphate (paper III) was determined with x-ray diffraction (XRD). The regularly arranged crystal lattice produce characteristic peaks for each specific material (brushite, monetite, pyrophosphate, etc.), in the XRD spectra. The relative amounts of each phase can be calculated using a mathematical analysis of the spectra, Rietveld refinement. In paper I and II the phase composition was analyzed with automated software, Profex. Since the pyrophosphate precipitant in paper III was amorphous, chemical analysis with Fourier transformed infra- red spectroscopy (FTIR) was used to detect absorbance in the infra-red range of light. Each material has a characteristic absorbance/transmittance pattern, and the observed patterns were compared to recently published patterns[138, 139].

59 Imaging The surface and morphology of monetite scaffolds (paper II) and pyrophosphate precipitant (paper III) were characterized by scanning electron microscopy (SEM). Samples were sputter coated with gold/palladium, approximately 5-10nm in thickness. An electron beam excited the sample to emit electrons (secondary and back scattered). The pattern of detected electrons were used to generate an image, as seen in papers II and III.

Porosity measurement Brushite or monetite cylinders were polished and samples without obvious defects were measured with calipers. The dry weight of each sample was recorded and the apparent density (ρa) was calculated by equation 5, where m is the recorded dry mass and v is the calculated volume of the cylinder. The samples were then ground to a fine powder and the density of the powder was determined with a helium pycnometer. The obtained density value, known as skeletal density (ρs), represents the density of the individual crystals that compose the cement. If the cylinder is composed entirely of brushite/ monetite there will be no porosity, and the skeletal and apparent densities will be identical. The total porosity (ϕ) was determined by dividing the apparent density by the hypothetical density of a non-porous scaffold (ρs) (equation 6). ! �! = Equation 5 !

!! � = (1 − ) * 100% Equation 6 !!

The microporosity was calculated for control cylinders that contained no PEG, and, therefore, no theoretical macroporosity. The macroporosity was then determined by subtracting the total porosity value for control scaffolds from the total porosity values for each group. This approach assumes that the microporosity remains the same regardless of the amount of macroporosity, between groups. MicroCT measurements were conducted to confirm the amount of macroporosity (<100um pore size). A pixel size of 10um was used when identifying pores larger than 100um.

Mechanical testing Brushite or monetite cylinders were compressed (uniaxial) on an AGS-X mechanical tester, at a crosshead speed of 1 mm per minute. For ten- sile/screw pull testing the orthopedic screw was inserted into cement (Figure

60 25 left image) or Sawbone/cement (Figure 25 right image) and allowed to set in cement for 24 hours at >90% humidity. The sample was loaded such that the screw was held straight and fixed to prevent sliding, and the screw head was pulled away from the Sawbone at 1mm per minute. A metal plate held the Sawbone sample in a fixed location, while the screw was pulled. In groups without cortical fixation the hole was large (2 cm) to prevent cement from contacting the metal plate. In cortical fixation groups the hole through which the screw was pulled was narrow to ensure that the cement and Saw- bone came into contact with the cortical shell as the screw was pulled out (Figure 20C).

Figure 25. The testing setup and forces applied to determine the strength of the screw/cement bond (A), and the testing rig used to pull screws from Sawbone (B). Reproduced from paper IV.

Drug release The release rate of Rebamipide was determined by high pressure liquid chromatography (HPLC) in paper I. Leach extracts from the cement were collected, as described in paper I, sterile filtered (0.2 um) to remove debris, then loaded on a Waters 2965 HPLC machine. Rebamipide was dissolved in a mobile phase of 33.25% acetonitrile, 64.75% water, 1% acetic acid and 1% methanol, and retained in a YMC Triart C18 column. After an approximate 3 minute retention time, Rebamipide exited the column and was detected by UV light (246nm). The amount of Rebamipide was calculated by comparing the total area under the curve (AUC) of the Rebamipide peak to standards with known concentrations. The observed rate of drug release was close to pure fickian diffusion (Higuchi model fit R = 0.99), though the release exponent (0.56-0.65) suggests that (Table 4) erosion may also contribute to the faster than expected release.

61 Biological testing Cell testing was conducted by either indirect or direct contact with cement, or pyrophosphate. Cements (0.8g) containing Rebamipide were leached into 10mL of phosphate buffered saline, with complete replacement of the saline at every time point. The concentration of Rebamipide in the leach was determined by HPLC and the leach was diluted as described in paper I. Cells testing with direct contact was conducted in paper II and III. In paper II cells were seeded directly onto the scaffolds, and in paper III calcium pyrophosphate precipitant formed in the media and came into contact with the apical (upper) surface of the cell membrane. The following cell types were use for biological testing: osteoblasts (MC3T3, Saos-2), macrophages (Raw264.7), and stem cells (1013A). The MC3T3 are early stage, mouse pre-osteoblasts that originate from the skull (calavarial). These cells can be pushed towards a mature osteoblast pheno- type by adding ascorbic acid (which causes collagen production) or culturing on collagen, and supplying a phosphate source such as beta-glycerol phosphate. Saos-2 are late stage, mature human osteosarcoma (cancer) cells. Both osteoblast cell types produce a collagen matrix, alkaline phosphatase, and mineralize the matrix under osteogenic conditions. The mouse macrophage cell line (Raw264.7) can respond to inflammatory stimuli (phorbol myristate acetate (PMA) and lipopolysaccharide (LPS)) by producing reactive oxidative species (ROS). The human induced pluripotent stem cell line (hiPSC) 1013A was originally derived from human dermal cells that were repro- grammed back to an embryonic stage. These pluripotent cells were differentiated towards mesenchymal stem cell fate, prior to culture. Cell testing included viability, differentiation, and gene expression. Cell viability was tested with alamar blue (paper I, III) and presto blue (paper II). Both viability reagents use active ingredients that change color when reduced by the respiration machinery of the cell (electron transport). The amount of cells correlate with the absorbance/emission produced by the reduced product, with the assumption that the metabolic rate, and rate of alamar blue or presto blue reduction, is roughly equivalent between all cells and experiments. Viability measurements are commonly expressed in terms of percentage survival, where the absorbance/emission readings are normalized to either experimental and control groups, or between later and initial time points. The morphology and survival of adherent cells was further examined with Live/Dead stain, which includes a dye that is converted to an intracellularly restricted fluorescent dye by metabolic enzymes (Live), and a second dye that enters the nucleus if the nuclear membrane is compromised (propid- ium iodide, Dead). Osteoblast differentiation was identified by ALP activity (paper I, III) and gene expression (paper II, III). The activity of alkaline phosphatase, a key indicator of osteoblast differentiation, was determined by first isolating the

62 intracellular ALP in a lysis buffer, followed by freeze and thaw cycles to ensure the cells were ruptured and all intracellular ALP was liberated. A synthetic substrate for ALP, p-nitrophenylphosphate, was incubated with ALP to produce a yellow product (4-nitrophenol, 4NP) that absorbs at 405 nm. The ALP activity is directly correlated with the amount of 4NP that is produced, per unit of time. The ALP readings were normalized to the amount of protein for each sample because the cell number, and protein content) varies between samples. Gene expression was determined by real time PCR. RNA was isolated and purified with Trizol, quantitated, and reverse transcribed to cDNA. The DNA was then quantitated over multiple cycles of amplification and normalized to a housekeeping gene (beta-Actin).

Chemiluminescence Reactive oxygen species production was measured with a luminol based chemiluminescence assay. Macrophages were stimulated to produce reactive oxygen species with phorbol myristate acetate (1uM), after which luminol (50uM) was added and the emission at 492 nm was measured kinetically. The radicals produced by macrophages oxidize luminol and photons are produced as luminol is converted to a peroxide form. The total amount of photons produced (area under the curve), over an hour incubation time, was calculated and expressed, for each treatment group and normalized to control (no Rebamipide or leach).

63 Svensk sammanfattning

Varje år drabbas miljontals människor av benbrott eller skelettrelaterade sjukdomar. När benvävnaden inte kan läka ordentligt kan olika biomaterial tex kalciumfosfatcement användas för att underlätta läkningsprocessen och för att stabilisera det skadade benet. Det kan exempelvis användas för att fylla hålrum efter borttagna tumörer eller för att förstärka ben vid fixering av benbrott. Den kemiska sammansättningen av kalciumfosfatcement är nästan identisk med naturligt benmineral, vilket betyder att cementet kan integreras väl med benet och att det över tid kan ersättas med ny benvävnad. Trots att detta syntetiska material har utmärkta biologiska egenskaper så är kalicumfosfatcement begränsade till användning i icke-lastbärande tillämpningar, som exempelvis i armen eller skallbenet, eftersom cementen vanligtvis är ganska svaga och sköra. En optimerad version av kalciumfosfatcement med högre hållfasthet har tidigare utvecklats i vårt laboratorium. Cementet kan nå en styrka på upp till 90 MPa i tryckbelastning, vilket tyder på att det kan vara lämpligt för vissa lastbärande tillämpningar. Målet med denna avhandling var att utvärdera lämpligheten av detta optimerade cementet för lastbärande applikationer samt att undersöka hur de modifierade fysikaliska och biologiska egenskaperna hos denna version av kalciumfosfatcement skiljer sig från de ursprung- liga cementen med lägre styrka. I artikel I undersöktes frisättningen av ett läkemedel, som kan stimulera benläkning, från cementet. När cementpulver blandas med vätska och börjar härda bildas porer (hålrum som upptas av vatten under cementhärdningsre- aktionen). Mängden porer avgör hur lätt vätska kan komma in i cementet och därmed också hur snabbt ett läkemedlet kan ta sig ut. Läkemedelsubstansen (Rebamipide) blandades in i cementet och frisattes sedan när detta placera- des i en vätska. Eftersom detta cement också har en för cement låg andel porer undersöktes även hastigheten med vilken läkemedlet frisattes från cementet. Den huvudsakliga upptäckten i artikel I var att det optimerade kalciumfosfatcementet kunde frisätta läkemedel inom ett biologiskt aktivt dose- ringsspann med en kontinuerlig frisättning på upp till 28 dagar. I artikel II undersöktes interaktionen mellan mänskliga stamceller och ytan på cementet. Eftersom benmineral och kalicumfosfatcement kemiskt liknar varandra så kan benceller och stamceller lätt fästa, växa till och diffe- rentiera på kalciumfosfatcement. Cementets yta ändras dock beroende på kemi samt tillverknings- och processtegen, vilket gjorde det nödvändigt att

64 studera hur celler skulle reagera på det optimerade cementet. Den främsta upptäckten i artikel II var att mänskliga stamceller växte till och differentie- rade på cementet i upp till 7 veckor och uttryckte benspecifika gener jämför- bara med celler som hade odlats på bovin benvävnad. Under den biologiska testprocessen som användes i artikel I, exponerades benceller för urlakade extrakt från cementet. Det visade sig att extraktet i hög utsträckning stimulerade alkalisk fosfatasaktivitet. Alkaliskt fosfatas (ALP) är ett nyckelenzym som möjliggör bildandet av nytt benmineral (kal- ciumfosfat). ALP är därmed en allmänt använd indikator för minseralisering in vitro. I artikel III följde vi upp denna observation genom att undersöka om det fanns bioaktiva komponenter i urlakningsextraktet. Det är redan känt att kalcium- och fosfatjoner stimulerar benceller (osteoblaster), därför foku- serade vi på det oorganisa additivet pyrofosfat. Pyrofosfat används i många typer av kalciumbaserade cement för att sakta ned reaktionshastigheten och på så sätt förbättra de fysikaliska egenskaperna, såsom tryckhållfasthet. I cementet binder pyrofosfat till kalcium och bildar kalciumpyrofosfat, vilket bromsar den kemiska reaktionen och leder till en reducerad kristallstorlek. Intressant nog använder människokroppen pyrofosfat för samma ändamål: för att inhibera mängden av kristalltillväxt av benmineral. I artikel III expo- nerade vi osteoblaster från mus och människa för pyrofosfat. Det huvudsakliga målet med artikel III var att fastställa om pyrofosfat påverkade tillväxt- hastigheten eller differentieringshastigheten av omogna osteoblaster till mogna osteoblaster. Det visade sig att pyrofosfat stimulerade ALP- aktiviteten i osteoblaster utan att påverka tillväxthastigheten. Intressant nog ökade också genuttrycket för vissa viktiga proteiner i benvävnadsmatrisen när de utsattes för pyrofosfat, vilket inkluderande kollagen I (COL1), ostepontin (OPN) och osteokalcin (OCN). Resultaten från artikel III tyder på att pyrofosfat inte bara inhiberar mineralisering utan även kan stimulera proteiner som är involverade i tillväxten av ny benvävnad. Artikel IV undersökte cementets prestanda i en lastbärande modell. Orto- pediska benskruvar används vanligtvis för att förankra metallplattor, som nyttjas vid stabiliseringen av benbrott. Skruven kan lossna från vävnaden om en hög kraft appliceras eller om benet nära skruven är av ”dålig” kvalité. Trycket på gränssnittet mellan benet och skruvgängan kan överstiga styrkan hos benet. När benet kring skruvgängan fallerar kan skruven börja röra sig och så småningom lossna från benet. Även måttlig belastning kan så små- ningom leda till att skruven lossnar som ett resultat av utmattning och mik- rosprickor som ackumuleras vid gränsytan mellan skruv och ben. Cement kan injiceras in i gränsytan mellan skruv och ben för att öka kontaktytan mellan skruvgängorna och benvävnaden. Cementering av ortopediska skruvar är känt för att signifikant öka den kraft som behövs för att dra ut skruven ur benvävnaden och därför förväntas sannolikheten för att ska lossna in vivo att minska. Det huvudsakliga målet med artikel IV var att i en modell med cementerade ortopediska skruvar bestämma huruvida en ökning av cement-

65 hållfastheten resulterade i en högre kraft med vilken skruven hölls på plats. Eftersom egenskaperna hos ben varierar kraftigt med skillnader i bentäthet och trabekulär tjocklek baserat på ålder, kön, fysiologisk plats, etc användes en syntetisk motsvarighet till ben, ett polyuretanskum kallat ”Sawbone”, istället för benvävnad. Skruvar för trabekulärt ben skruvades in i Sawbone och den maximala kraften som behövdes för att dra ut skruven registrerades. De huvudsakliga resultaten från artikel IV var att a) porositeten och tryck- hållfastheten hos kalciumfosfatcementet bestämmer kraften som krävs för att cementerade benskruvar ska lossna, b) kraften som leder till att skruvar lossnar ur benet korrelerar tillräckligt starkt med cementets porositet och tryck- hållfasthet att den kan förutsägas för en känd volym cement och porositet, c) kvaliteten och tätheten i det omgivande benet påverkar kraften som leder till att skruvar lossnar. Denna avhandlings slutsatser är: a) det optimerade kalciumfosfatcement med lägre porositet och högre hållfasthet effektivt kan frisätta terapeutiska medel på ett kontinuerligt och kontrollerat sätt i upp till 28 dagar, b) cementet stöder vidhäftning, expansion och differentiering av mänskliga stamceller, c) urlakade extrakt från cementen, särskilt pyrofosfat, främjar benspecifika genuttryck i osteoblaster från mus och människa och d) ortopediska skruvar cementerade med det optimeratde cement kräver betydligt högre kraft innan skruven lossnar från benet i jämförelse med ett cement med lägre hållfasthet.

66 Acknowlegements

First and foremost, I want to thank my supervisors. Håkan and Marjam, you have both given me the greatest opportunity of my life. Every day I get to solve puzzles, riddles, and problems. To walk the line between absolute dis- sapointment (also known as cell culture experimentation) and exhilarating success, and to discover something unknown, is true happiness. Your guidance styles are perfectly complimentary. Håkan, from you I have learned how to ask material science questions, to see the big picture, to plan and to always have a “strategy”. You have included me on high impact projects, given me a chance to collaborate and exchange knowledge in New York, Spain, London, and Switzerland. Marjam, from you I have learned how to ask the biological questions, how to test every variable, but also to maximize the results while minimizing the amount of work needed. You have always encouraged me, and counted me as part of your family. Thank you both, thank you, thank you. I also want to thank the many people in the MIM and engineering research group who have helped me. Ingrid, Johanna, Eric, Alejandro, Oscar, Lee, Luimar, Charlotte, Wei, Caroline, Cecilia, Marjam, Maryam, Miriam, Shiuli, Tao, Song, Carl, Xi, Sara G., Gry, David, Ernesto, Victoria, Petra, Celine, Jun, Le, Viviana, Sandeep, Sujit, Maruthi, Lisa, Robin, Gemma, and Sara R., Ingrid R., and everyone else. You have each made the lab a brighter, better place. Whenever I need help or to talk you guys have been there. Thanks Marjam, Ingrid, Caroline, Cecilia and Gry for making my Swedish summary possible, I could not do it without you!!!! Thank you Johanna, Ingrid, Ale, and Shiuli for training and guidance when I first joined the lab. Thank you Shiuli, Marjam, Håkan, Celine, Sujit, and Caroline for always taking the time to discuss problems and experiments! Thanks Caroline, Jun, Celine, Petra, Charlotte, Ernesto, Sara G., Sujit, and Wei- you have all helped me out in challenging situations when I couldn’t finish an experiment or technique alone (XPS, feeding cells while on vacation, etc.). Thank you to Sandeep, Sujit, Maruthi, Viviana, Shiuli, Marjam, and now Gry for all the cell lab conversations- I have learned a kilo-ton from you guys! A special thank you to my collaborators and funding sources: Thomas Lind for your insight and guidance with molecular biology; Giuseppe M. dePeppo and Martina Sladkova at New York Stem Cell Foundation for host- ing me, sharing your knowledge and the good times we had in New York; Biodesign FP7; Swedish Science Foundation; Swedish Research Council;

67 and Philip and Gerard for always asking questions that no one else has an- swered yet, the amazing opportunities, and for teaching me that there isn’t a problem that can’t be solved with either the screw or the glue. A special second round of thanks to Håkan, Philip and Gerard- we have worked very closely together and I am very lucky to have gotten to know each of you better than would have otherwise been possible through typical research projects. The future is genuinely bright, and I can’t wait to start the next phase of research and life with you all. A big thank you to my previous supervisors and teachers. Dr. Yang, and Dr. Tang- thank you especially for your patience, training and guidance; Dr. Randolph Christen, Michael Ger- hold, Sara Axeford, Edward Modafferi: thank you for your training as my first supervisors. Everything I have, or will, accomplish is a result of the training, resources and efforts you have freely given to me. I will never for- get that. Thank you to my previous supervisor Veena; you took a chance on me, and the training, mentorship, and opportunity you provided has led me to where I am today. Thank you to previous teachers and mentors, as well, that have made a huge impact on my life: Steve Larsson (triple thank you Mr. Larsson), Linda Dillard, Tom Scanlon, Katherine Tanney, Milton Saier, George Redmond. And to my closes friends, Nadja, James, Johan and Lucie (JoLu), Kaspar, Charlotte, Sharcus (Sharn and Marcus), Le, my blueberry and smultron plants, and the book “Structure and chemistry of the apatites and other calcium orthophosphates”: I wish I had more time to spend- I so cherish the times we have had, and the things I have learned from you. I look forward to spending more time with each of you after the defense, and once my library card is renewed. We are talking weekly dinners guys, weekly! Ashwin and Sapna, you have always supported us and given such good advice. Congratu- lations on the baby! We wish you all the success and good fortunate one can handle. Ashleigh, it has been so much fun planning Shiuli’s defense party and hanging out. I look forward to the next 4 years together! Thank you to my first, best, true friends: Brian David, Van N., Vittoric/John C., Tyler, Veronica W., Carmen C., Sam F. A huge thank you to my 2nd family, Jan, Fari, Marjam, Sascha, Mina, Nora, Milo, and Dante. Jan and Fari, you have helped me time after time, and been so wonderful to me. Thank you for all the times we have spent together, including house hunting, Persian new year dinners, the physallis plant! Sascha, you're such a cool guy, I have enjoyed the time we have had together and looking for the many times to come. Mina, thank you for sup- porting me in my plant endeavours and your love for scifi tv shows and movies. Nora, Milo, and Dante, you guys have grown so fast, I am so happy that I get to spend time with you, you really make me laugh and smile. Lastly, a giant thank you to my family. To my parents, thank you for be- lieving in me and guiding me, for teaching me how to survive, and how to live. To my mother Isabell, you have sacrificed endlessly so that I could

68 have the opportunities you did not have. Each day I think upon this as I en- joy the many privileges I now have in my life, like living in a safe neighbor- hood, having food in the refrigerator and a safe place to sleep each night, an excellent education, and having a life full of love. Thank you, I love you Mom. To my Ma and Pa, you have welcomed me as a son, and I am so grateful that I am part of the Pujari clan. It is the best thing that has ever happened to me. You guys have always supported me, taught me how to live life, I wish and hope that I can one day repay you for all the kindness and love you have given me. I wish we had more time together, but as soon as you move to Sweden we'll hang out every day! I love you both Ma and Pa! To my sisters, Nicole, Cassandra, and Ratul- I am so proud of you guys! Nicole and Cassandra, life gave you quite a lot of lemons, and yet you have not given up. Never give up. Thank you for teaching me the value of drink- ing lemonade. Ratul, you have had many challenges to face and have always stayed strong. I am so impressed at how far you have come in such a short time. I am waiting to go to the fields to practice some archery soon. Thanks for being the best sisters in the world! Michael (2), you have a wonderful name, and you've been a great brother. I hope we can spend more time together, I can definitely use some help when it comes to repairs and renova- tions around the house ☺. Destini, Excel and Eber, I am so happy to have you guys as nieces and nephews, you always surprise and inspire me. I can’t wait to see you guys grow up! Shiuli, my amazing wife, you have changed my life completely, and for the better. I am so grateful we got the chance to work together, and I miss it everyday. I look forward to the day when you are fired and must rejoin me at the lab. In all seriousness though, you give me the strength to face any challenge, I look forward to spending everyday with you until the day I die. I love you more than french fries, always. You are my proudest accomplish- ment and, though we have not reached an agreement on which one of us married up more, I think we can both agree that we both married up.

69 References

[1] Langer R, Vacanti JP. Tissue engineering. Science. 1993;260:920-6. [2] Sheikh ZJ, M.A.; Hamdan, N.; Hashmi, R. . Bone Regeneration Using Bone Morphogenetic Proteins and Various Biomaterial Carriers. . Materials 2015:1778-816. [3] Boskey AL. Bone composition: relationship to bone fragility and antiosteoporotic drug effects. BoneKEy reports. 2013;2:447. [4] Chai YC, Carlier A, Bolander J, Roberts SJ, Geris L, Schrooten J, et al. Current views on calcium phosphate osteogenicity and the translation into effective bone regeneration strategies. Acta biomaterialia. 2012;8:3876-87. [5] Laurencin C, Khan Y, El-Amin SF. Bone graft substitutes. Expert review of medical devices. 2006;3:49-57. [6] Ginebra MP, Traykova T, Planell JA. Calcium phosphate cements as bone drug delivery systems: a review. Journal of controlled release : official journal of the Controlled Release Society. 2006;113:102-10. [7] Tamimi F, Sheikh Z, Barralet J. Dicalcium phosphate cements: brushite and monetite. Acta biomaterialia. 2012;8:474-87. [8] P. BAYHvBCH. Osteoinductive biomaterials: current knowledge of properties, experimental models and biological mechanisms. Eur Cell Mater. 2011;21:407- 29. [9] Wang Y, Yu X, Baker C, Murphy WL, McDevitt TC. Mineral particles modulate osteo-chondrogenic differentiation of embryonic stem cell aggregates. Acta biomaterialia. 2016;29:42-51. [10] Wang X, Xu S, Zhou S, Xu W, Leary M, Choong P, et al. Topological design and additive manufacturing of porous metals for bone scaffolds and orthopaedic implants: A review. Biomaterials. 2016;83:127-41. [11] B. ZXvdMKHSNWAHdC. Chondrocytes transdifferentiate into osteoblasts in endochondral bone during development, postnatal growth and fracture healing in mice. PLoS Genet. 2014;10. [12] Einhorn TA, Gerstenfeld LC. Fracture healing: mechanisms and interventions. Nature reviews Rheumatology. 2015;11:45-54. [13] Thompson Z, Miclau T, Hu D, Helms JA. A model for intramembranous ossification during fracture healing. Journal of orthopaedic research : official publication of the Orthopaedic Research Society. 2002;20:1091-8. [14] Kaji H, Sugimoto T, Kanatani M, Chihara K. High extracellular calcium stimulates osteoclast-like cell formation and bone-resorbing activity in the presence of osteoblastic cells. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 1996;11:912- 20. [15] Raggatt LJ, Partridge NC. Cellular and molecular mechanisms of bone remodeling. The Journal of biological chemistry. 2010;285:25103-8.

70 [16] Oftadeh R, Perez-Viloria M, Villa-Camacho JC, Vaziri A, Nazarian A. Biomechanics and mechanobiology of trabecular bone: a review. Journal of biomechanical engineering. 2015;137. [17] Novitskaya EC, PY; Hamed, E; Jun, L; Lubarda, V, A.; Jasiuk, I; McKittrick, J. Recent advances on the measurement and calculation of the elastic moduli of cortical and trabecular bone: A review. Theoretical and Applied Mechanics. 2011;38:209-97. [18] Basel D, Steiner RD. Osteogenesis imperfecta: recent findings shed new light on this once well-understood condition. Genetics in medicine : official journal of the American College of Medical Genetics. 2009;11:375-85. [19] Ham A. The Function of Bone As A Calcium Reservoir With A Consideration of the Cellular Pictures Seen in Resorption With Particular Reference to the Significance of Osteoclasts. The Angle Orthodontist. 1932;2:142-57. [20] Polo-Corrales L, Latorre-Esteves M, Ramirez-Vick JE. Scaffold design for bone regeneration. Journal of nanoscience and nanotechnology. 2014;14:15-56. [21] Zappone B, Thurner PJ, Adams J, Fantner GE, Hansma PK. Effect of Ca2+ ions on the adhesion and mechanical properties of adsorbed layers of human osteopontin. Biophysical journal. 2008;95:2939-50. [22] Kindt JH, Thurner PJ, Lauer ME, Bosma BL, Schitter G, Fantner GE, et al. In situ observation of fluoride-ion-induced hydroxyapatite-collagen detachment on bone fracture surfaces by atomic force microscopy. Nanotechnology. 2007;18:135102. [23] Fantner GE, Adams J, Turner P, Thurner PJ, Fisher LW, Hansma PK. Nanoscale ion mediated networks in bone: osteopontin can repeatedly dissipate large amounts of energy. Nano letters. 2007;7:2491-8. [24] Ferreira AM, Gentile P, Chiono V, Ciardelli G. Collagen for bone tissue regeneration. Acta biomaterialia. 2012;8:3191-200. [25] Mizuno M, Fujisawa R, Kuboki Y. Type I collagen-induced osteoblastic differentiation of bone-marrow cells mediated by collagen-alpha2beta1 integrin interaction. Journal of cellular physiology. 2000;184:207-13. [26] Fantner GE, Hassenkam T, Kindt JH, Weaver JC, Birkedal H, Pechenik L, et al. Sacrificial bonds and hidden length dissipate energy as mineralized fibrils separate during bone fracture. Nature materials. 2005;4:612-6. [27] Fantner GE, Oroudjev E, Schitter G, Golde LS, Thurner P, Finch MM, et al. Sacrificial bonds and hidden length: unraveling molecular mesostructures in tough materials. Biophysical journal. 2006;90:1411-8. [28] Hansma PK, Fantner GE, Kindt JH, Thurner PJ, Schitter G, Turner PJ, et al. Sacrificial bonds in the interfibrillar matrix of bone. Journal of musculoskeletal & neuronal interactions. 2005;5:313-5. [29] Thibault MM, Hoemann CD, Buschmann MD. Fibronectin, vitronectin, and collagen I induce chemotaxis and haptotaxis of human and rabbit mesenchymal stem cells in a standardized transmembrane assay. Stem cells and development. 2007;16:489-502. [30] Giachelli CM, Steitz S. Osteopontin: a versatile regulator of inflammation and biomineralization. Matrix biology : journal of the International Society for Matrix Biology. 2000;19:615-22. [31] Steitz SA, Speer MY, McKee MD, Liaw L, Almeida M, Yang H, et al. Osteopontin inhibits mineral deposition and promotes regression of ectopic calcification. The American journal of pathology. 2002;161:2035-46.

71 [32] Addison WN, Masica DL, Gray JJ, McKee MD. Phosphorylation-dependent inhibition of mineralization by osteopontin ASARM peptides is regulated by PHEX cleavage. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2010;25:695-705. [33] Chen Y, Bal BS, Gorski JP. Calcium and collagen binding properties of osteopontin, bone sialoprotein, and bone acidic glycoprotein-75 from bone. The Journal of biological chemistry. 1992;267:24871-8. [34] Zou C, Song G, Luo Q, Yuan L, Yang L. Mesenchymal stem cells require integrin beta1 for directed migration induced by osteopontin in vitro. In vitro cellular & developmental biology Animal. 2011;47:241-50. [35] Gericke A, Qin C, Spevak L, Fujimoto Y, Butler WT, Sorensen ES, et al. Importance of phosphorylation for osteopontin regulation of biomineralization. Calcified tissue international. 2005;77:45-54. [36] Engleman VW, Nickols GA, Ross FP, Horton MA, Griggs DW, Settle SL, et al. A peptidomimetic antagonist of the alpha(v)beta3 integrin inhibits bone resorption in vitro and prevents osteoporosis in vivo. The Journal of clinical investigation. 1997;99:2284-92. [37] Ross FP, Chappel J, Alvarez JI, Sander D, Butler WT, Farach-Carson MC, et al. Interactions between the bone matrix proteins osteopontin and bone sialoprotein and the osteoclast integrin alpha v beta 3 potentiate bone resorption. The Journal of biological chemistry. 1993;268:9901-7. [38] Nishimichi N, Hayashita-Kinoh H, Chen C, Matsuda H, Sheppard D, Yokosaki Y. Osteopontin undergoes polymerization in vivo and gains chemotactic activity for neutrophils mediated by integrin alpha9beta1. The Journal of biological chemistry. 2011;286:11170-8. [39] Ihara H, Denhardt DT, Furuya K, Yamashita T, Muguruma Y, Tsuji K, et al. Parathyroid hormone-induced bone resorption does not occur in the absence of osteopontin. The Journal of biological chemistry. 2001;276:13065-71. [40] Ishijima M, Rittling SR, Yamashita T, Tsuji K, Kurosawa H, Nifuji A, et al. Enhancement of osteoclastic bone resorption and suppression of osteoblastic bone formation in response to reduced mechanical stress do not occur in the absence of osteopontin. The Journal of experimental medicine. 2001;193:399- 404. [41] Asou Y, Rittling SR, Yoshitake H, Tsuji K, Shinomiya K, Nifuji A, et al. Osteopontin facilitates angiogenesis, accumulation of osteoclasts, and resorption in ectopic bone. Endocrinology. 2001;142:1325-32. [42] Thurner PJ, Chen CG, Ionova-Martin S, Sun L, Harman A, Porter A, et al. Osteopontin deficiency increases bone fragility but preserves bone mass. Bone. 2010;46:1564-73. [43] Pietak AM, Sayer M. Functional atomic force microscopy investigation of osteopontin affinity for silicon stabilized tricalcium phosphate bioceramic surfaces. Biomaterials. 2006;27:3-14. [44] Bellahcene A, Bonjean K, Fohr B, Fedarko NS, Robey FA, Young MF, et al. Bone sialoprotein mediates human endothelial cell attachment and migration and promotes angiogenesis. Circulation research. 2000;86:885-91. [45] Adams J, Fantner GE, Fisher LW, Hansma PK. Molecular energy dissipation in nanoscale networks of Dentin Matrix Protein 1 is strongly dependent on ion valence. Nanotechnology. 2008;19:384008. [46] Malaval L, Wade-Gueye NM, Boudiffa M, Fei J, Zirngibl R, Chen F, et al. Bone sialoprotein plays a functional role in bone formation and osteoclastogenesis. The Journal of experimental medicine. 2008;205:1145-53.

72 [47] Golub EE. Role of matrix vesicles in biomineralization. Biochimica et biophysica acta. 2009;1790:1592-8. [48] Millan JL. Alkaline Phosphatases : Structure, substrate specificity and functional relatedness to other members of a large superfamily of enzymes. Purinergic signalling. 2006;2:335-41. [49] Shih YR, Hwang Y, Phadke A, Kang H, Hwang NS, Caro EJ, et al. Calcium phosphate-bearing matrices induce osteogenic differentiation of stem cells through adenosine signaling. Proceedings of the National Academy of Sciences of the United States of America. 2014;111:990-5. [50] Polewski MD, Johnson KA, Foster M, Millan JL, Terkeltaub R. Inorganic pyrophosphatase induces type I collagen in osteoblasts. Bone. 2010;46:81-90. [51] Okada S, Kiyama T, Sato E, Tanaka Y, Oizumi T, Kuroishi T, et al. Inhibition of phosphate transporters ameliorates the inflammatory and necrotic side effects of the nitrogen-containing bisphosphonate zoledronate in mice. The Tohoku journal of experimental medicine. 2013;231:145-58. [52] Silver IA, Murrills RJ, Etherington DJ. Microelectrode studies on the acid microenvironment beneath adherent macrophages and osteoclasts. Experimental cell research. 1988;175:266-76. [53] Bose S, Fielding G, Tarafder S, Bandyopadhyay A. Understanding of dopant- induced osteogenesis and angiogenesis in calcium phosphate ceramics. Trends in biotechnology. 2013;31:594-605. [54] Schack LM, Noack S, Winkler R, Wissmann G, Behrens P, Wellmann M, et al. The Phosphate Source Influences Gene Expression and Quality of Mineralization during In Vitro Osteogenic Differentiation of Human Mesenchymal Stem Cells. PloS one. 2013;8:e65943. [55] Muller P, Bulnheim U, Diener A, Luthen F, Teller M, Klinkenberg ED, et al. Calcium phosphate surfaces promote osteogenic differentiation of mesenchymal stem cells. Journal of cellular and molecular medicine. 2008;12:281-91. [56] Kang H, Shih YR, Varghese S. Biomineralized matrices dominate soluble cues to direct osteogenic differentiation of human mesenchymal stem cells through adenosine signaling. Biomacromolecules. 2015;16:1050-61. [57] Chai YC, Roberts SJ, Schrooten J, Luyten FP. Probing the osteoinductive effect of calcium phosphate by using an in vitro biomimetic model. Tissue engineering Part A. 2011;17:1083-97. [58] Suzuki E, Ochiai-Shino H, Aoki H, Onodera S, Saito A, Saito A, et al. Akt activation is required for TGF-beta1-induced osteoblast differentiation of MC3T3-E1 pre-osteoblasts. PloS one. 2014;9:e112566. [59] Camalier CE, Yi M, Yu LR, Hood BL, Conrads KA, Lee YJ, et al. An integrated understanding of the physiological response to elevated extracellular phosphate. Journal of cellular physiology. 2013;228:1536-50. [60] Beck GR, Jr., Zerler B, Moran E. Phosphate is a specific signal for induction of osteopontin gene expression. Proceedings of the National Academy of Sciences of the United States of America. 2000;97:8352-7. [61] Choudhary S, Wadhwa S, Raisz LG, Alander C, Pilbeam CC. Extracellular calcium is a potent inducer of cyclo-oxygenase-2 in murine osteoblasts through an ERK signaling pathway. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2003;18:1813- 24. [62] Chattopadhyay N, Yano S, Tfelt-Hansen J, Rooney P, Kanuparthi D, Bandyopadhyay S, et al. Mitogenic action of calcium-sensing receptor on rat calvarial osteoblasts. Endocrinology. 2004;145:3451-62.

73 [63] Viti F, Landini M, Mezzelani A, Petecchia L, Milanesi L, Scaglione S. Osteogenic Differentiation of MSC through Calcium Signaling Activation: Transcriptomics and Functional Analysis. PloS one. 2016;11:e0148173. [64] Nakade O, Takahashi K, Takuma T, Aoki T, Kaku T. Effect of extracellular calcium on the gene expression of bone morphogenetic protein-2 and -4 of normal human bone cells. Journal of bone and mineral metabolism. 2001;19:13- 9. [65] Dvorak MM, Siddiqua A, Ward DT, Carter DH, Dallas SL, Nemeth EF, et al. Physiological changes in extracellular calcium concentration directly control osteoblast function in the absence of calciotropic hormones. Proceedings of the National Academy of Sciences of the United States of America. 2004;101:5140- 5. [66] Eapen A, Sundivakkam P, Song Y, Ravindran S, Ramachandran A, Tiruppathi C, et al. Calcium-mediated stress kinase activation by DMP1 promotes osteoblast differentiation. The Journal of biological chemistry. 2010;285:36339- 51. [67] Sun Y, Zeng XR, Wenger L, Cheung HS. Basic calcium phosphate crystals stimulate the endocytotic activity of cells--inhibition by anti-calcification agents. Biochemical and biophysical research communications. 2003;312:1053- 9. [68] Halverson PB, Greene A, Cheung HS. Intracellular calcium responses to basic calcium phosphate crystals in fibroblasts. Osteoarthritis and cartilage. 1998;6:324-9. [69] Pujari-Palmer M, Pujari-Palmer S, Lu X, Lind T, Melhus H, Engstrand T, et al. Pyrophosphate Stimulates Differentiation, Matrix Gene Expression and Alkaline Phosphatase Activity in Osteoblasts. PloS one. 2016;11:e0163530. [70] Christoffersen MS, N; Zunicb, T;Christoffersen, J. Kinetics of dissolution of triclinic calcium pyrophosphate dihydrate crystals. Journal of Crystal Growth 1999;203:234-43. [71] Lomashvili KA, Garg P, Narisawa S, Millan JL, O'Neill WC. Upregulation of alkaline phosphatase and pyrophosphate hydrolysis: potential mechanism for uremic vascular calcification. Kidney international. 2008;73:1024-30. [72] Harmey D, Hessle L, Narisawa S, Johnson KA, Terkeltaub R, Millan JL. Concerted regulation of inorganic pyrophosphate and osteopontin by akp2, enpp1, and ank: an integrated model of the pathogenesis of mineralization disorders. The American journal of pathology. 2004;164:1199-209. [73] Nudelman F, Pieterse K, George A, Bomans PH, Friedrich H, Brylka LJ, et al. The role of collagen in bone apatite formation in the presence of hydroxyapatite nucleation inhibitors. Nature materials. 2010;9:1004-9. [74] Hodge AJ. Molecular models illustrating the possible distributions of 'holes' in simple systematically staggered arrays of type I collagen molecules in native- type fibrils. Connective tissue research. 1989;21:137-47. [75] Stamov DP, T. Structure and function of ECM-inspired composite collagen type I scaffolds. Soft Matter. 2012;8:10200-12. [76] Anderson HC. Matrix vesicles and calcification. Current rheumatology reports. 2003;5:222-6. [77] Hasegawa TY, T.; Tsuchiya, E.; Hongo, H.; Tsuboi, K.; Kudo, A.; Abe, M.; Yoshida, T.; Nagai, T.; Khadiza, N.; Yokoyama, A.; Oda, K.; Ozawa, O.; Luiz de Freitas, P.; Li, M; Amizuka, N. Ultrastructural and biochemical aspects of matrix vesicle-mediated mineralization. Japanese Dental Science Review. 2016.

74 [78] Ciancaglini P, Simao AM, Camolezi FL, Millan JL, Pizauro JM. Contribution of matrix vesicles and alkaline phosphatase to ectopic bone formation. Brazilian journal of medical and biological research = Revista brasileira de pesquisas medicas e biologicas. 2006;39:603-10. [79] Thouverey C, Bechkoff G, Pikula S, Buchet R. Inorganic pyrophosphate as a regulator of hydroxyapatite or calcium pyrophosphate dihydrate mineral deposition by matrix vesicles. Osteoarthritis and cartilage. 2009;17:64-72. [80] Mehta M, Schmidt-Bleek K, Duda GN, Mooney DJ. Biomaterial delivery of morphogens to mimic the natural healing cascade in bone. Advanced drug delivery reviews. 2012;64:1257-76. [81] Shieh SJ, Terada S, Vacanti JP. Tissue engineering auricular reconstruction: in vitro and in vivo studies. Biomaterials. 2004;25:1545-57. [82] Zein NN, Hanouneh IA, Bishop PD, Samaan M, Eghtesad B, Quintini C, et al. Three-dimensional print of a liver for preoperative planning in living donor liver transplantation. Liver transplantation : official publication of the American Association for the Study of Liver Diseases and the International Liver Transplantation Society. 2013;19:1304-10. [83] Huebsch N, Mooney DJ. Inspiration and application in the evolution of biomaterials. Nature. 2009;462:426-32. [84] Wengst UA, M. The mechanical efficiency of natural materials. Philosophical Magazine. 2004;84. [85] Marsh CH, R. Bone porosity and the use of mercury intrusion porosimetry in bone diagenesis studies. Archaeomerry. 1999;41. [86] Karageorgiou V, Kaplan D. Porosity of 3D biomaterial scaffolds and osteogenesis. Biomaterials. 2005;26:5474-91. [87] Marenzana M, Arnett TR. The key role of the blood supply to bone. Bone research. 2013;1:203-15. [88] Lovett M, Lee K, Edwards A, Kaplan DL. Vascularization strategies for tissue engineering. Tissue engineering Part B, Reviews. 2009;15:353-70. [89] Itala AI, Ylanen HO, Ekholm C, Karlsson KH, Aro HT. Pore diameter of more than 100 microm is not requisite for bone ingrowth in rabbits. Journal of biomedical materials research. 2001;58:679-83. [90] Lee JT, Chow KL, Wang K, Tsang WH. Is macroporosity absolutely required for preliminary in vitro bone biomaterial study? A comparison between porous materials and flat materials. Journal of functional biomaterials. 2011;2:308-37. [91] Descamps MD, T;. Monchaua, F.; Luc, J.; Hardouinc, P.; Horneza, J.; Lerichea, A. Manufacture of macroporous β-tricalcium phosphate bioceramics. Journal of the European Ceramic Society. 2008;28. [92] Bohner M, Baumgart F. Theoretical model to determine the effects of geometrical factors on the resorption of calcium phosphate bone substitutes. Biomaterials. 2004;25:3569-82. [93] De Groot K. Effect of porosity and physicochemical properties on the stability, resorption, and strength of calcium phosphate ceramics. Annals of the New York Academy of Sciences. 1988;523:227-33. [94] Lian CZ, Y.; Beechama, S. The relationship between porosity and strength for porous concrete. Construction and Building Materials. 2011;25. [95] Bignon AC, J.; Chevalier, J.;Fantozzi, G.;Carret, J.;Chavassieux, P.;Boivin, G.;Melin, M.;Hartmann, D. Effect of micro- and macroporosity of bone substitutes on their mechanical properties and cellular response. Journal of Materials Science: Materials in Medicine. 2003;14.

75 [96] Kendall KH, A.;Birchall, J.; Pratt, P.;Proctor, B.; Jefferis, S. The Relation between Porosity, Microstructure and Strength, and the Approach to Advanced Cement-Based Materials. Philosophical Transactions of the Royal Society of London Series A, Mathematical and Physical Sciences. 1983;310. [97] Dorozhkin S. Calcium Orthophosphates as Bioceramics: State of the Art. Journal of functional biomaterials. 2010;1. [98] Chen CH, Z.; Yuan, W.; Li, J.; Chenga, X.; Chia, R. Pressure effecting on morphology of hydroxyapatite crystals in homogeneous system. CrystEngComm. 2011;13. [99] Pujari-Palmer S, Chen S, Rubino S, Weng H, Xia W, Engqvist H, et al. In vivo and in vitro evaluation of hydroxyapatite nanoparticle morphology on the acute inflammatory response. Biomaterials. 2016;90:1-11. [100] Cunniffe GM, Curtin CM, Thompson EM, Dickson GR, O'Brien FJ. Content- Dependent Osteogenic Response of Nanohydroxyapatite: An in Vitro and in Vivo Assessment within Collagen-Based Scaffolds. ACS applied materials & interfaces. 2016;8:23477-88. [101] Hesaraki S, Nazarian H, Pourbaghi-Masouleh M, Borhan S. Comparative study of mesenchymal stem cells osteogenic differentiation on low-temperature biomineralized nanocrystalline carbonated hydroxyapatite and sintered hydroxyapatite. Journal of biomedical materials research Part B, Applied biomaterials. 2014;102:108-18. [102] Unosson JE, H. Development of a Resorbable Calcium Phosphate Cement with Load Bearing Capacity. Bioceramics Development and Applications. 2014. [103] Davies E, Muller KH, Wong WC, Pickard CJ, Reid DG, Skepper JN, et al. Citrate bridges between mineral platelets in bone. Proceedings of the National Academy of Sciences of the United States of America. 2014;111:E1354-63. [104] Guimaraes LF, Fidalgo TK, Menezes GC, Primo LG, Costa e Silva-Filho F. Effects of citric acid on cultured human osteoblastic cells. Oral surgery, oral medicine, oral pathology, oral radiology, and endodontics. 2010;110:665-9. [105] Grover LM, Wright AJ, Gbureck U, Bolarinwa A, Song J, Liu Y, et al. The effect of amorphous pyrophosphate on calcium phosphate cement resorption and bone generation. Biomaterials. 2013;34:6631-7. [106] Rasmusson L, Abtahi J. Bisphosphonate associated osteonecrosis of the jaw: an update on pathophysiology, risk factors, and treatment. International journal of dentistry. 2014;2014:471035. [107] Hofmann MP, Mohammed AR, Perrie Y, Gbureck U, Barralet JE. High- strength resorbable brushite bone cement with controlled drug-releasing capabilities. Acta biomaterialia. 2009;5:43-9. [108] Brown CJ, Sinclair RA, Day A, Hess B, Procter P. An approximate model for cancellous bone screw fixation. Computer methods in biomechanics and biomedical engineering. 2013;16:443-50. [109] Lee SK, K. Current Concepts in Management of Tibia Plateau Fracture. Journal of the Korean Fracture Society. 2014;27. [110] Collinge C, Merk B, Lautenschlager EP. Mechanical evaluation of fracture fixation augmented with tricalcium phosphate bone cement in a porous osteoporotic cancellous bone model. Journal of orthopaedic trauma. 2007;21:124-8. [111] Eriksson F, Mattsson P, Larsson S. The effect of augmentation with resorbable or conventional bone cement on the holding strength for femoral neck fracture devices. Journal of orthopaedic trauma. 2002;16:302-10.

76 [112] Seo JH, Ju CI, Kim SW, Kim JK, Shin H. Clinical efficacy of bone cement augmented screw fixation for the severe osteoporotic spine. Korean Journal of Spine. 2012;9:79-84. [113] Grund J, Korber M, Bodmeier R. Predictability of drug release from water- insoluble polymeric matrix tablets. European journal of pharmaceutics and biopharmaceutics : official journal of Arbeitsgemeinschaft fur Pharmazeutische Verfahrenstechnik eV. 2013;85:650-5. [114] Peppas NA. 1. Commentary on an exponential model for the analysis of drug delivery: Original research article: a simple equation for description of solute release: I II. Fickian and non-Fickian release from non-swellable devices in the form of slabs, spheres, cylinders or discs, 1987. Journal of controlled release : official journal of the Controlled Release Society. 2014;190:31-2. [115] Levander G. A study of bone regeneration. Surg Gynecol Obstet. 1938;67:705-14. [116] Urist MR. Bone: formation by autoinduction. Science. 1965;150:893-9. [117] Miron RJ, Zhang YF. Osteoinduction: a review of old concepts with new standards. Journal of dental research. 2012;91:736-44. [118] Liu C, Chen CW, Ducheyne P. In vitro surface reaction layer formation and dissolution of calcium phosphate cement-bioactive glass composites. Biomedical materials. 2008;3:034111. [119] Nilsson M, Wang JS, Wielanek L, Tanner KE, Lidgren L. Biodegradation and biocompatability of a calcium sulphate-hydroxyapatite bone substitute. The Journal of bone and joint surgery British volume. 2004;86:120-5. [120] Lin L, Chow KL, Leng Y. Study of hydroxyapatite osteoinductivity with an osteogenic differentiation of mesenchymal stem cells. Journal of biomedical materials research Part A. 2009;89:326-35. [121] Gustavsson J, Ginebra MP, Engel E, Planell J. Ion reactivity of calcium- deficient hydroxyapatite in standard cell culture media. Acta biomaterialia. 2011;7:4242-52. [122] Olkowski R, Kaszczewski P, Czechowska J, Siek D, Pijocha D, Zima A, et al. Cytocompatibility of the selected calcium phosphate based bone cements: comparative study in human cell culture. Journal of materials science Materials in medicine. 2015;26:270. [123] Przekora AC, J.;Pijocha, D.;Ślósarczyk, A.;Ginalska, G. Do novel cement- type biomaterials reveal ion reactivity that affects cell viability in vitro? 9. 20139;3. [124] Byrne DL, D.; Planell, J.; Kelly, D.; Prendergast, P. Simulation of tissue differentiation in a scaffold as a function of porosity, Young's modulus and dissolution rate: Application of mechanobiological models in tissue engineering. Biomaterials. 2007;28. [125] Hoppe A, Guldal NS, Boccaccini AR. A review of the biological response to ionic dissolution products from bioactive glasses and glass-ceramics. Biomaterials. 2011;32:2757-74. [126] Seebeck J, Goldhahn J, Stadele H, Messmer P, Morlock MM, Schneider E. Effect of cortical thickness and cancellous bone density on the holding strength of internal fixator screws. Journal of orthopaedic research : official publication of the Orthopaedic Research Society. 2004;22:1237-42. [127] Pujari-Palmer M, Pujari-Palmer S, Engqvist H, Karlsson Ott M. Rebamipide delivered by brushite cement enhances osteoblast and macrophage proliferation. PloS one. 2015;10:e0128324.

77 [128] Sladkova M, Palmer M, Ohman C, Alhaddad RJ, Esmael A, Engqvist H, et al. Fabrication of macroporous cement scaffolds using PEG particles: In vitro evaluation with induced pluripotent stem cell-derived mesenchymal progenitors. Materials science & engineering C, Materials for biological applications. 2016;69:640-52. [129] Sun WH, Tsuji S, Tsujii M, Gunawan ES, Kawai N, Kimura A, et al. Induction of cyclooxygenase-2 in rat gastric mucosa by rebamipide, a mucoprotective agent. The Journal of pharmacology and experimental therapeutics. 2000;295:447-52. [130] Lai Y, Zhong W, Yu T, Xia ZS, Li JY, Ouyang H, et al. Rebamipide Promotes the Regeneration of Aspirin-Induced Small-Intestine Mucosal Injury through Accumulation of beta-Catenin. PloS one. 2015;10:e0132031. [131] Nanke Y, Kobashigawa T, Yago T, Kawamoto M, Yamanaka H, Kotake S. Rebamipide, an Amino Acid Analog of 2(1H)-Quinolinone, Inhibits the Formation of Human Osteoclasts. BioMed research international. 2016;2016:6824719. [132] Naito Y, Yoshikawa T. Rebamipide: a gastrointestinal protective drug with pleiotropic activities. Expert review of gastroenterology & hepatology. 2010;4:261-70. [133] Engstrand T, Kihlstrom L, Lundgren K, Trobos M, Engqvist H, Thomsen P. Bioceramic Implant Induces Bone Healing of Cranial Defects. Plastic and reconstructive surgery Global open. 2015;3:e491. [134] Engstrand T, Kihlstrom L, Neovius E, Skogh AC, Lundgren TK, Jacobsson H, et al. Development of a bioactive implant for repair and potential healing of cranial defects. Journal of neurosurgery. 2014;120:273-7. [135] Kurien T, Pearson RG, Scammell BE. Bone graft substitutes currently available in orthopaedic practice: the evidence for their use. The bone & joint journal. 2013;95-B:583-97. [136] Ajaxon I, Ohman Magi C, Persson C. Compressive fatigue properties of an acidic calcium phosphate cement-effect of phase composition. Journal of materials science Materials in medicine. 2017;28:41. [137] Yoshida H, Faust A, Wilckens J, Kitagawa M, Fetto J, Chao EY. Three- dimensional dynamic hip contact area and pressure distribution during activities of daily living. Journal of biomechanics. 2006;39:1996-2004. [138] Gras PR, C.; Marsan, O.; Sarda, S.; Combes, C. Synthesis and characterisation of hydrated calcium pyrophosphate phases of biological interest. European Journal of Inorganic Chemistry. 2013. [139] van Leeuwen YV, K.; Kegel, W. Morphology of colloidal metal pyrophosphate salts. RSC Advances. 2012;2.

Acta Universitatis Upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1502 Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science and Technology, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology. (Prior to January, 2005, the series was published under the title “Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology”.)

ACTA UNIVERSITATIS UPSALIENSIS Distribution: publications.uu.se UPPSALA urn:nbn:se:uu:diva-319495 2017