Characterization of biological mineralization in vitro

Characterization of biological mineralization in vitro

Karakterisatie van biologische mineralisatie in vitro (met een samenvatting in het Nederlands)

Proefschrift

ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof. dr. W.H. Gispen,

ingevolge het besluit van het college voor promoties in het openbaar te verdedigen

op donderdag 14 december 2006 des middags te 12.45 uur

door

Leonie Fleur Arjanne Huitema geboren op 30 juli 1978, te Soest, Nederland

Promotoren: Prof. Dr. J.B. Helms

Prof. Dr. A. Barneveld Co-promotoren: Dr. A.B.Vaandrager

Dr. P.R. van Weeren Dr. C.H.A. van de Lest

Dit proefschrift werd mede mogelijk gemaakt met financiële steun van de J.E. Jurriaanse Stichting.

Contents Chapter 1 What triggers cell-mediated mineralization? 1 Chapter 2 Rapid cell-mediated large mineralizing structures in

a serum free cell culture system 29 Appendix A Subcellular localization of mineral crystals in

ATDC5 cells 49 Appendix B Inorganic phosphate does not induce a major effect on the

cytosolic free calcium concentration in ATDC5 cells 57 Chapter 3 Soluble factors released by ATDC5 cells affect the

formation of calcium phosphate crystals 63 Chapter 4 The nitric oxide donor sodium nitroprusside inhibits

mineralization in ATDC5 cells 79 Chapter 5 Iron ions derived from the nitric oxide donor sodium

nitroprusside inhibit mineralization 95 Chapter 6 Summarizing discussion 111 Samenvatting in het Nederlands 121 Abbreviations 127 Dankwoord 129 Curriculum Vitae 132

Printed by Ridderprint, Ridderkerk ISBN-10: 90-393-4393-4 ISBN-13: 978-90-393-4393-7

Chapter 1

General introduction

Based on

What triggers cell-mediated mineralization?

Leonie F.A. Huitema1,2 and Arie B. Vaandrager1

1 Department of Biochemistry and Cell Biology, 2 Department of Equine Sciences, Faculty of Veterinary Medicine, and Graduate School Animal Health.

Front Biosci. 2007; 12; 2631-2645

Chapter 1

2

General introduction into cell-mediated mineralization Mineralization is an essential requirement for normal skeletal development, which is generally accomplished through the function of two cell types, osteoblasts and chondrocytes [1]. Soft tissues do not mineralize under normal conditions but, under certain pathological conditions some tissues like articular cartilage and cardiovascular tissues are prone to mineralization (Table 1) [2;3]. Mineralization of articular cartilage contributes to significant morbidity because of its association with joint inflammation and worsening of the progression of osteoarthritis [4]. Articular cartilage calcification occurs in association with aging, degenerative joint disease (e.g. osteoarthritis), some genetic disorders and various metabolic disorders [4-7]. Similarly, arterial calcification occurs with advanced age, atherosclerosis, metabolic disorders, including end stage renal disease and diabetes mellitus, and some genetic disorders [8]. Arterial calcification contributes to hypertension and increased risks of cardiovascular events, leading to morbidity and mortality [9]. While pathological mineralization has long been considered to result from physiochemical precipitation of calcium and phosphate, recent studies have provided evidence that soft tissue mineralization is a regulated process, which has many similarities with bone formation [10;11]. For now it is unclear why soft tissues have the tendency to mineralize. Therefore, the purpose of this review is to discuss the components involved in cell-mediated mineralization. 1. Different types of mineralization 1.1 Intramembranous ossification Clavicles and bones in the regions of the craniofacial skeleton develop by intramembranous ossification, the principle of which is shown in Fig. 1, panel 1. During intramembranous ossification, osteoblasts originating from mesenchymal cell condensations are responsible for bone and matrix deposition [12]. During this mesenchymal cell condensation, high densities of mesenchymal cells aggregate and start to produce an extracellular matrix [13]. Subsequently, mesenchymal cells acquire the typical columnar shape of osteoblasts, increase the synthesis of alkaline phosphatase (APase), and begin to secrete bone matrix, which consists mainly of type I collagen (Fig. 1, panel 1A). Osteoblasts then generate inorganic phosphate (Pi) by an increase of APase, which is necessary for bone formation [14;15]. Numerous ossification centres then develop and eventually fuse [12]. Finally, osteoblasts undergo apoptosis (~ 70%) or terminally differentiate to form osteocytes (~ 30%), which become entrapped in the mineralized bone matrix [14;16;17]. Osteocytes are in fact osteoblasts buried within the mineralized bone matrix (Fig. 1, panel 1B).

Table 1. Forms of pathological mineralization Organ Disease Clinical pattern Etiology References Joints Tendonitis Calcification of the tendon Inflammation, injury [146;147] overload Arthritis Calcification of articular cartilage Inflammation, injury [9;38;148] overload, aging Bursitis Calcification of bursal walls Inflammation, overload [149-151] Aging Bone spurs Bony projections that grow along Repetitive trauma [152;153] the edges of the joints (osteophytes) Heart Heart valve calcification Calcification in/around myofibroblasts Aging, inherited, infection [154] Blood vessels Atherosclerosis Calcification/bone formation in/around Inflammation, injury [44;155] vascular smooth muscle cells Aging, high blood pressure Muscles Myositis ossificans Bone formation in the muscles Injury [156] Lung Metastatic pulmonary Calcification in alveolar septa End stage renal disease [157;158] Calcification Pulmonary ossification Bone formation in alveolar compartments Inflammation [158] Skin Cutaneous ossification Bone formation in the skin Injury, inflammation [159] Eyes Cataracts Calcification of the lens Aging [160] Kidney Kidney stones Calcium oxalate or calcium Pi stones Inherited, mineral imbalance [161] Brain Bilateral Striopallidodentate Symmetric calcification of the basal ganglia Aging, inherited, infection [162;163] calcinosis Spinal cord Neurogenic heterotopic Bone formation in spinal cord Injury [164] ossification Tumor Osteosarcoma Bone cancer that may metastasize elsewhere Unknown [165]

Chapter 1

4

Once embedded, they cease their ossification activity and communicate with each other and with cells at the bone surface via a meshwork of cell processes that run through canaliculi in the bone matrix [16]. A key regulator of osteoblast differentiation and function is the transcription factor low core binding protein alpha 1 (Cbfa1) [12;18-21]. Cbfa1 is targeted to the promotors of several bone proteins, such as osteocalcin, bone sialoprotein, APase and type I collagen [18]. No bone tissue is formed in Cbfa1 null mice, although a complete cartilaginous skeleton is formed, indicating that Cbfa1 is not essential for chondrogenesis [19]. To maintain bone structure and skeletal growth, remodeling is necessary. The remodeling of bone that has been formed by either intramembranous ossification or endochondral ossification (see below) consists of a strict coupling of bone resorption by osteoclasts and bone formation by osteoblasts that continue throughout life, with a positive balance during growth and with a negative balance during ageing. Osteoclasts are multinucleated giant cells formed from hemopoietic precursors of the monocyte and macrophage series [22]. They attach to the bone surface by sealing a resorbing compartment that they acidify by secreting H+ ions. This will then dissolve the bone mineral, and subsequently expose the organic matrix to proteolytic enzymes that degrade it [22]. 1.2 Endochondral ossification Vertebrate long bones form through a process called endochondral ossification, in which a cartilage template produced by chondrocytes, is formed and replaced by bone (Fig. 1, panel 2) [12]. During the process of endochondral ossification, mesenchymal cell condensation results in differentiation into chondrocytes [13]. Chondrocytes then produce a framework of skeletal cartilage template that will subsequently be replaced by bone [12]. It has been suggested that this cartilage template in long bones is formed because the mineralized extracellular matrix of bone limits interstitial growth. To achieve rapid and directional growth in an organism, an intermediate structure (cartilage) that can function under high load, and at the same time generate space for new bone formation, would then be very helpful [23]. Generally 4 zones are recognized during endochondral ossification Fig. 1, panel 2A-D). The upper zone is the resting zone (Fig. 1, panel 2A), chondrocytes are small and dormant, and mainly type II collagen is produced. Subsequently, in the proliferative zone chondrocytes start to proliferate in vertical columns (Fig. 1, panel 2B). The next zone is the hypertrophic zone (Fig. 1, panel 2C), where the chondrocytes start to enlarge, produce type X collagen, and increase the synthesis of APase. Finally, hypertrophic chondrocytes induce mineralization (Fig. 1, panel 2D), which is correlated with increased levels of Pi in the mineralization zone [24-26]. Furthermore, terminally differentiated chondrocytes in the growth plate are deleted from the cartilage by programmed cell death [23]. In


5

this final zone, osteoprogenitor cells and hematopoietic stem cells arrive via the newly formed blood vessels, which penetrate through the transverse septa of mineralizing hypertrophic chondrocytes. Osteoprogenitor cells beneath the site of vascular invasion differentiate into osteoblasts that aggregate on the surface of the calcified cartilage and deposit bone matrix (osteoid) Fig. 1 panel, 2E). Bone resorbing osteoclasts start to remodel the newly formed bone [27]. The transcription factor SOX9 is mainly expressed in resting and proliferating chondrocytes, while it is switched off in hypertrophic chondrocytes [28]. SOX9 is required for expression of cartilage specific extracellular matrix components, such as collagen types II, IX and XI [12;28]. In humans, mutations in SOX9 result in a skeletal malformation syndrome called campomelic dysplasia [29;30]. No chondrocyte specific markers are expressed in SOX9 null cells in mouse chimeras [31]. Interestingly, Cbfa1 expression is detected in hypertrophic chondrocytes, which suggests that hypertrophic chondrocytes may undergo a phenotypic transition towards osteoblast phenotype [12]. Furthermore, hypertrophic chondrocytes in the femur and humerus are absent in Cbfa1 null mice [21]. The hypothesis of chondrocytes transdifferentiating into osteoblasts is supported by microscopic examinations of cells present at the chondro-osseous junction in the growth plate [32-35]. Based on this observation, it is speculated that there are two subpopulations of chondrocytes, one which becomes apoptotic, while the other population transdifferentiates into osteoblasts [32]. 1.3 Pathological calcification Soft tissues in organisms, like soft tissue structures in joints, skin, heart, blood vessels, kidneys, muscles, lungs, etc. do not mineralize under normal conditions [3]. However, as shown in Table 1, under certain pathological conditions some organs mineralize. This is also called pathological mineralization or dystrophic calcification [36]. In particular vascular smooth muscle cells in blood vessels are prone to mineralization (Fig. 1 panel 3A-B). Recent studies have provided evidence that pathological vascular calcification is a regulated process, which has many similarities with bone formation [3;11;37;38]. Expression of a variety of bone-associated proteins has been found in atherosclerotic plaques [39;40]. Furthermore, it has been reported that 10 to 30% of the cells in a smooth muscle cell culture system undergo a dramatic phenotypic transition in the presence of relatively high Pi (> 2 mM). This is characterized by the loss of smooth muscle cell lineage marker expression and upregulation of genes related to the osteogenic phenotype [10;41-43]. This suggests that vascular smooth muscle cells undergo phenotypic transition to osteochondroprogenitor-like cells. Formation of complete bone tissue and bone marrow has been demonstrated in calcified arteries, a phenomenon also called heterotopic (extraosseous) calcification [44;45]. Regulated

Chapter 1

6

changes in chondrocyte differentiation and viability characteristically seen in growth plate chondrocytes can also occur in mineralizing articular cartilage chondrocytes as a result of osteoarthritis [4].

Figure 1. Simplified schematic representation of different forms of cell mediated mineralization. Hatched area represent mineral. 1. Intramembranous ossification. A) Mesenchymal cells aggregate and form nodules (ossification centre), where the cells will differentiate into osteoblasts, B) Bone is formed and remodeled by osteoblasts forming bone matrix and by osteoclasts resorbing bone. Osteocytes entrapped in the mineralized matrix. 2. In endochondral ossification 4 zones are recognized in the epiphyseal growth plate. A) resting zone: chondrocytes are small and dormant, mainly type II collagen is produced, B) proliferative zone: chondrocytes start to proliferate in vertical columns, type II collagen is produced, C) Hypertrophic zone: chondrocytes enlarge and start to produce type X collagen, D) Mineralizing zone: hypertrophic chondrocytes start to mineralize, produce matrix vesicles and finally die, Type I collagen is now produced, E) Osteoblasts grow on the mineralization sheath and produce bone and type I collagen. Some osteoblasts become entrapped in the mineralized matrix and become osteocytes. Osteoclasts remodel newly formed bone. 3. Pathological mineralization (e.g. atherosclerosis). A) Vascular smooth muscle cells in a blood vessel, B) As a result of an atherosclerotic lesion, vascular smooth muscle cells differentiate into osteoblast-like cells and start to induce mineralization.


7

Figure 2. Schematic diagram of factors proposed to be involved in cell mediated mineralization. Bold arrows indicate that an increase of phosphate (Pi) rather than calcium (Ca) induces cell-mediated mineralization. (1) Tissue fluid contains Ca and Pi. An increase of Pi may induce mineralization. (2) Cells can also generate a local increase of Pi, when they increase APase-levels. The increased Pi induces: (3) the production of MVs, which contain APase and also generate Pi, (4) cell death, which results in the formation of apoptotic bodies, (5) production of nucleating proteins, which are secreted in the extracellular matrix. (6) To control mineralization, the cells constitutively produce mineralization inhibitors. (7) Tissue fluid also contains mineralization inhibitors, principally fetuin that originates from serum.

Chapter 1

8

2. Mechanism of cell-mediated mineral deposition To induce cell-mediated mineralization, the organism has to create an environment with a local increase of calcium and/or Pi, and subsequently organize the nucleation of these ions in an ordered fashion. It has been reported that Pi levels increase considerably from the proliferative to the hypertrophic zone in the growth plate [25;26]. Patients with end-stage renal disease (ESRD) develop vascular calcification, which is correlated with an increased serum Pi concentration that typically exceeds 2.0 mM (normal level: 1.4 mM) [10;46]. In addition, APase, which cleaves phospho-compounds to Pi, is highly increased in mineralization competent cells like osteoblasts, hypertrophic chondrocytes and mineralizing vascular smooth muscle cells. The role of APase in mineralization is essential, since APase deficient mice show impaired skeletal mineralization [47]. However, the identity of the physiological organic Pi substrate is not defined. There is much debate regarding the biochemical mechanisms that initiate mineralization subsequent to the increase in calcium and/or Pi. Fig. 2 summarizes the principal factors that are believed to play a role in the process of tissue mineralization. One theory proposes that mineralization is initiated within matrix vesicles (MVs) [48]. A second, not mutually exclusive, theory proposes that Pi induces apoptosis, and that apoptotic bodies nucleate crystals composed of calcium and Pi [23;49]. A third theory suggests that mineralization is mediated by certain non-collagenous proteins (NCPs), which associate with the extracellular matrix [50-52]. Next to proteins that induce mineralization, organisms also actively inhibit mineralization by secreting specific proteins and removal of an inhibitor may induce mineralization as well (Fig. 2) [53]. 2.1 The role of matrix vesicles It has been hypothesized that cell-mediated mineralization is induced within matrix vesicles (MVs). MVs are cell-derived extracellular membrane enclosed particles, about 0.1- 1 µm in size [48]. In a mineralization inducing environment, they are proposed to bud off the outer cell plasma membrane in a polarized fashion to the longitudinal septal matrix in the growth plate and to the newly formed osteoid under the mineral facing surfaces of osteoblasts in bone [48;54]. MVs have been isolated from mineralizing odontoblasts, osteoblasts, chondrocytes and vascular smooth muscle cells [55-57]. This is generally performed after a crude collagenase digestion (typically collagenase 500 U/ml at 37°C for 3 hours). After gentle vortexing, MVs are then harvested by differential centrifugation. To this end, the collagenase digest is centrifuged at 13,000 x g for 20 minutes, and the resulting cells and cell debris are discarded. Subsequently, the supernatant is spun at 100,000 x g for 1 hour, which results in a pellet that contains MVs [58-62].


9

Many studies have been directed at the elucidation of the mechanism of MVs mineralization. In those studies, MVs isolated from tissues or cell culture systems were induced to calcify. This is generally performed by incubating isolated MVs in a so called synthetic cartilage lymph medium which is a physiological buffer containing approximately 2 mM calcium and 1.5 mM Pi [47;56;60;63-66]. From these studies it was concluded that MVs have to be mineralization competent to nucleate calcium Pi, because isolated MVs from non-mineralizing tissues do not calcify in synthetic cartilage lymph [59;67]. Unlike the mineralization competent MVs, these latter MVs did not express Annexin V and had a lower APase activity [47;59]. These observations suggest that not all MVs are equivalent and that only mineralizing tissue can produce mineralization competent MVs. These studies are complicated by the presence of vesicles derived from apoptotic cells (apoptotic bodies) in the matrix, which have properties similar but not identical to MVs (see next section) [61]. Several reports suggest that MVs do not contain crystals at the time of their release from the cell, but that the first crystalline mineral appears after the MV has been immobilized in selected areas of the collagen matrix [48]. The mineral crystal is then formed by concentrating calcium and Pi at the inner leaflet of the vesicle membrane, which has been reported to be enriched in phosphatidylserine (PS) [68]. Once the mineral has reached a certain size it ruptures the vesicle membrane and contributes to the extracellular matrix [48]. The mechanisms by which the crystals break down or penetrate through the membrane are not fully understood. Because MVs are enclosed by a membrane, channel proteins are required to mediate the influx of mineral ions into these particles. Uptake of Pi is critical for the formation of minerals within MVs. It has been shown that the sodium dependent type III Pi transporter Glvr-1 is mainly expressed in the growth plate by early hypertrophic chondrocytes, which are the MVs producing cells [69]. MVs isolated from chicken epiphyseal cartilage have been shown to contain a sodium dependent Pi transport system [70]. Other studies report that MV mineralization is not strictly sodium dependent, suggesting the presence of other Pi transporters as well [63]. MVs also have the potential to cleave phospho-compounds to Pi, resulting in a local increase of Pi concentration, as APase has been shown to be enriched in the membrane of MVs [48;71]. Annexin II, V and VI have been reported to mediate the influx of calcium into the vesicles, by forming hexamers in the PS enriched MV membrane [72;73]. Furthermore, Annexin V binds type II and X collagen and this interaction has been shown to stimulate its calcium channel activity [60]. The role of Annexin V was further established through suppression by siRNA, which resulted in inhibited mineralization, while overexpression stimulated mineralization [74;75]. Annexin expression has also been found to be increased in hypertrophic chondrocytes from

Chapter 1

10

the growth plate [76]. Annexin V knockout mice did not show an impaired skeletal phenotype and neither were the in vitro calcification properties of the isolated annexin (-/-) chondrocytes significantly impaired [77]. Possibly other members of the annexin family compensate the deficiency in annexin. 2.2 The role of cell death There is a strong correlation between mineralization and cell death. Especially pathological mineralization has often been associated with apoptotic or necrotic processes [11]. Furthermore, terminally differentiated chondrocytes in the growth plate are deleted from the cartilage by programmed cell death [23;36]. Pi, whose concentration is reported to increase from the proliferative to the hypertrophic zone in the growth plate, has been shown to be a potent apoptogen [78;79]. A Pi-induced intracellular effect was evidenced, since Na-Pi transporter inhibitors were shown to inhibit apoptosis in parallel with mineralization. It should be noted that Pi-induced cell death was strongly synergized by the extracellular calcium concentration [69;80;81]. A critical role of Pi in apoptosis was also established in mice affected with hypophosphatemia. These animals contain an expanded layer of late non hypertrophic chondrocytes in the growth plate, which is associated with a decrease in the number of apoptotic hypertrophic chondrocytes [82]. In contrast, patients affected with hyperphosphatemia show pathological mineralization, which correlated with an increase in cell death [83;84]. Pi is not the only agent that induces cell death during mineralization, since pathological mineralization also occurs in the absence of hyperphosphatemia, indicating the role of additional factors. In agreement with this, it has been reported that no calcification occurred in vessels with calcium and Pi concentrations of 1.8 mM and 3.8 mM respectively, but mechanical injury resulted in extensive calcification under these conditions [85]. Until now it is not clear how apoptosis contributes to mineralization. Possibly, dying or injured cells may become highly permeable to calcium and Pi, and may concentrate these ions beyond their solubility product, facilitating heterogeneous nucleation and crystal growth. It has been proposed that an early step in apoptosis is externalization of phosphatidylserine (PS) [86]. PS has been shown to have a high affinity for calcium and may act as a nucleator for calcium Pi crystal formation [36;87]. However, it has been reported that mineral-PS interactions can retard crystal growth [88]. This suggests that although PS has the capacity to nucleate calcium Pi in cells undergoing apoptosis, other factors, probably produced by living cells, are necessary to induce crystal growth. It has also been suggested that apoptotic bodies derived from dying cells may act as nucleating mineralization centres in a similar way as described for MVs. Apoptotic bodies isolated from cell culture systems have been shown to precipitate calcium Pi when incubated in a


11

synthetic cartilage lymph [56;89-91]. The hydroxyapatite precipitation was less in apoptotic bodies when compared to MVs [56]. This indicates that MVs have a stronger capacity to induce crystal growth. This was supported by Kirsch et al. (2003) who reported that apoptotic bodies do not contain APase and the calcium channel forming annexins II, V and VI [61]. Inhibiting apoptosis with a general caspase inhibitor has been shown to inhibit mineralization in cell culture systems by approximately 40%, indicating a role of apoptosis in mineralization, but also a role of factors other than apoptosis in the process of mineralization [78;92]. Recently, it has been proposed that mineralizing hypertrophic growth plate chondrocytes are not dying by a classical form of apoptosis, because, in contrast to in vitro cell culture systems, they do not produce apoptotic bodies in vivo [23;93]. Instead, it was speculated that they eliminate themselves by a process of autophagocytosis. This hypothesis is supported by ultramicroscopic examination of hypertrophic growth plate chondrocytes, showing that dying chondrocytes contain autophagic vacuoles (autophagosomes) and cell remnants that are blebbed off, indicative of autophagocytosis [23;93;94]. This specific form of cell death is also called chondroptosis [93]. Possibly, the blebs generated by these cells have mineralizing capacities. 2.3 The role of nucleating proteins Mineralization has also been proposed to be regulated by non-collagenous proteins (NCPs) found in the organic matrix of bone [52;95-99]. The NCPs are reported to constitute 5-10% of the total extracellular matrix and can be classified into four groups (Table 2). These groups include proteoglycans, glycoproteins, the γ carboxy glutamic acid (GLA)-containing proteins and the serum associated proteins [95;99;100]. Of these proteins mainly glycoproteins have been demonstrated to play a critical role in the initiation and growth of the calcium Pi mineral phase (Table 3). Glycoproteins are proteins that are modified posttranslationally by glycosylation, phosphorylation and sulfation. Some glycoproteins contain an RGD (Arg-Gly-Asp) sequence that interacts with the integrin receptor family. Glycoproteins that contain an RGD sequence are bone sialoprotein, BAG-75, dentin matrix protein-1, fibronectin, vitronectin, osteopontin and thromobospondin [99;101]. In vitro studies have shown that bone sialoprotein (BSP) can nucleate apatite crystals. BSP is an anionic phosphoprotein that is expressed almost exclusively in mineralized tissues [102;103]. It has been demonstrated that after treatment with organophosphate for 4-8 hours, BSP localizes to the extracellular matrix in osteoblastic cultures, well before the first appearance of apatite crystals [104-106]. This suggests that Pi (which is probably produced by APase) triggers BSP

Chapter 1

12

secretion into the extracellular matrix where it can subsequently nucleate calcium Pi in metastable solutions. It has also been reported that another noncollagenous bone matrix protein, bone acidic glycoprotein-75 (BAG-75) predicts the location of mineral nucleation, and possibly recruits BSP [51;52]. Purified BAG-75 can self-associate into supramolecular spherical complexes and sequesters millimolar quantities of Pi, which indicates that BAG-75 generates a localized Pi source for crystal nucleation reactions [107]. Interestingly, it has been proposed that BSP is associated with a population of vesicle-like structures (defined as crystal ghosts), which are 500-800 nm in size. However, BSP did not associate with the smaller 50-300 nm vesicle population [52]. An important role of BSP in mineralization has been further established by the observation that transfection of BSP into no mineralizing MC3T3-E1 subclones can restore their ability to form mineral deposits [51;108]. On the basis of this information BSP is likely to be involved in early mineral deposition. So far, the phenotype of BSP null mice is not known. Another NCP, dentin matrix protein-1 (DMP1), has been reported to facilitate hydroxyapatite growth [95;100]. A role of DMP1 in mineralization was suggested when it was shown to be mainly expressed during dentin mineralization, and later also in osteoblasts [109-111]. DMP1 null mice have a decreased mineral to matrix ratio in bones [112]. In an in vitro biomineralization model DMP1 has been shown to undergo a conformational change upon calcium binding and to subsequently assemble calcium Pi nuclei into ordered protein-mineral complexes. This results in an inhibiting effect on spontaneous calcium Pi precipitation. Thus, DMP1 could sequester and stabilize newly formed calcium Pi clusters [95]. Another in vitro study reported that DMP-1-induced crystal growth and proliferation is dependent on its degree of phosphorylation, because nonphosphorylated DMP1 acts as nucleators while the phosphorylated form inhibits nucleation [101]. Another NCP, fibronectin also has been shown to facilitate hydroxyapatite growth in the presence of a hydroxyapatite seed, and a close association between fibronectin and hydroxyapatite has been found in vivo [113;114]. Fibronectin, like DMP-1, has an inhibiting effect on spontaneous calcium Pi precipitation [95;114]. Therefore, it has been postulated that DMP-1 and fibronectin play a structural role in crystal growth, rather than a nucleating role.

biglycan BAG-75 protein S growth factors

Proteoglycans Glycoproteins with γ carboxy glutamic acid serum associated

decorin bone sialoprotein matrix GLA protein albumin versican osteopontin osteocalcin fetuin an RGD sequence (GLA)-containing proteins

Table 2. The non-collagenous proteins regulating mineralization can be classified into four groups [99].

thrombospondin

aggrecan dentin matrix protein-1

fibronectin

vitronectin

Chapter 1

14

2.4 The role of mineralization inhibitors In mammals, mineralization is generally controlled by two NCPs, fetuin and matrix GLA protein (Table 3) [53;115]. It has been proposed that the biological function of these proteins is to maintain high metastable blood calcium Pi levels and to inhibit unwanted (soft tissue) mineralization. Fetuin is synthesized in the liver and found in high concentrations in mammalian serum. Because of its high affinity to hydroxyapatite it is also found in bone and teeth [116-118]. Fetuin knockout mice spontaneously develop widespread soft tissue calcifications, including significant myocardial calcification [119]. In humans fetuin deficiency is associated with inflammation and vascular calcification [115;120]. In vitro studies demonstrate that fetuin inhibits precipitation of supersaturated solution of calcium and Pi by formation of a high molecular mass fetuin-mineral complex [121]. This complex prevents growth, aggregation and precipitation of calcium Pi[122]. Matrix GLA protein is an extracellular matrix protein that is generally expressed by chondrocytes and vascular smooth muscle cells [123]. Mice deficient in matrix GLA protein are normal at birth but develop severe calcification of all arteries (and cartilage) within weeks. These mice die at around 8 weeks of age, mostly due to a rupture of the aorta [123]. Interestingly, it has recently been reported that MVs isolated from vascular smooth muscle cells, that are induced to mineralize, contain fetuin and matrix GLA protein [56;124]. It has been speculated that this may be a defence mechanism of the cell to limit excessive mineralization [56;124]. The NCP osteopontin (OPN) has also been shown to control crystal growth [125]. In addition, Pi has been proposed to be a specific signal for upregulation of OPN gene expression, which supports its regulatory role during mineralization [15;126]. In agreement with this, gluteraldehyde fixed porcine aortic valves implanted into OPN null mice mineralized to a much greater extent than those implanted in wild type mice [127]. However, the inhibiting effect of OPN on mineralization is dependent on the extent of phosphorylation of OPN, because OPN dephosphorylated by APase does not inhibit mineralization [128-130]. In vivo, OPN is a protein that is normally found in mineralized tissue, but also in epithelial lining cells of numerous organs, and body fluids, including urine, saliva, milk and bile [129;131]. Next to inhibiting mineralization, OPN also regulates bone cell adhesion and osteoclast function in the skeleton [129]. When OPN attaches to osteoclasts, it stimulates the acidification of the local environment, which will allow for the dissolution of the mineral [129]. Another NCP, osteonectin (ON), also known as SPARC (secreted protein, acidic, rich in cysteine), is a calcium binding matrix protein found in many tissues undergoing remodeling [132]. Several in vitro studies have demonstrated that ON can inhibit crystal nucleation and retard crystal growth [133-135], although Hunter et al. (1996) found no effect [96]. An explanation for the different results may be


15

that different model systems were used, or the difference in ON concentrations tested. In vivo, ON deficient mice show an increased volume of adipose tissue and a decreased osteoblast and osteoclast number, resulting in osteopenia. This suggests that ON rather plays an important role in cell differentiation as well [132]. In vitro, similar results as with ON were obtained for the NCP osteocalcin (OC), which is also known as bone GLA protein (BGP) [96;99;133-135]. In addition, in vivo transgenic OC deficient mice demonstrate an increase in bone mass, which suggests that OC indeed limits mineralization [136]. 3. Scope of this thesis Pathological mineralization can have severe clinical consequences (Table 1). Articular cartilage calcification is one of the major degenerative diseases of the skeleton and leads to cartilage destruction, severe pain and joint stiffness and vascular calcification may lead to mortality [4;8;9]. Although many investigations greatly contributed to a better understanding of the mechanisms regulating cell-mediated mineralization, many questions remain about the mechanisms that trigger cell-mediated mineralization and how this process is regulated. It is still not clear whether one type of vesicles induces mineralization, or whether more types of vesicles are involved. This might possibly be different between different tissues. In addition, it is unclear where the mineral exactly nucleates. For example, does the first mineral develop extracellularly in vesicles? Or does it start intracellularly, which will then result in the formation of vesicles? Various proteins have been shown to be involved in mineralization, but it is unclear how these proteins are related mechanistically. Therefore, the aim of this study is to gain more insight in the processes that take place during the initiation of cell-mediated mineralization and the effectors involved in the process. In order to investigate cell-mediated mineralization in vitro, the chondrogenic mouse embryonal carcinoma-derived ATDC5 cell line was used (Fig. 3A). The ATDC5 cell line is a well characterized chondrogenic cell line (over 138 publications) shown to differentiate into chondrocytes by exposing the cells to 10 µg/ml insulin [137]. ATDC5 cells cultured for 21 days in differentiation medium form nodule like cell aggregates (Fig. 3B) in association with a dramatic elevation of APase activity and type X collagen expression, which suggests differentiation towards the hypertrophic phenotype [138]. Cbfa1 is elevated in ATDC5 cells prior to differentiation to the hypertrophic phenotype and treatment with antisense oligonucleotides for type I Cbfa1 severely reduced type X collagen expression in ATDC5 cells [139]. A subsequent medium switch to α-MEM has been shown to result in mineralization after another 14 days (Fig. 3C) [138;140]. MVs have been isolated from ATDC5 cells and were shown to take up Pi by an Na-dependent Pi

Chapter 1

16

transporter [141]. Furthermore, ATDC5 cells express Glvr-1 transcripts during various stages of their maturation, a process that has been shown to be sensitive to TGF-β1 [141;142]. The cell culture period required to induce mineralization can be reduced by addition of additional Pi, which increases the mineralization rate in the ATDC5 cell culture (Fig. 3D) [78;143]. Since we observed that mineralizing structures are strongly integrated in the extracellular matrix, making it difficult to investigate them, we decided to modify the ATDC5 cell culture system, as described in chapter 2. Mineralization was induced in the absence of serum, which is known to contain the mineralization inhibitor, fetuin [119;121]. This resulted in the formation of large mineralizing structures (LMS) in the medium in the sub millimeter range in size in a time period of 2 hours. LMS were shown to contain whole cells, which were embedded in hydroxyapatite and observed to have a stretched morphology. This may suggest that cell-mediated mineralization starts intracellularly. Therefore, an imaging study on cell-mediated mineralization in the presence of serum is described in appendix A to chapter 2 and an intracellular calcium imaging study is described in appendix B to chapter 2. To investigate whether ATDC5 cells themselves or factors released by ATDC5 cells nucleate and/or remodel calcium Pi in the absence of serum, the effect of conditioned medium from ATDC5 cells on the formation of apatite crystal is described in chapter 3. It was found that ATDC5 cells release constitutively soluble factors that affect the formation of calcium Pi crystals. Since an imbalance of mineralization may lead to pathological conditions, the ATDC5 cell culture system was used to test the effect of several agents implicated in bone growth and development as well as pathological mineralization, including the gaseous substance nitric oxide (NO) [144;145]. The effect of an NO donor drug, sodium nitroprusside (SNP) on cell mediated mineralization is described in chapter 4. It was found that SNP inhibits mineralization. However, the inhibition was not affected by inhibitors of guanylyl cyclase nor mimicked by a cGMP analog. Furthermore, sodium nitroprusside did not inhibit phosphate uptake nor inhibited apoptosis in the ATDC5 cells. Therefore, chapter 5 describes subsequent investigations into the effect of SNP on mineralization to elucidate the mechanism of action. In this chapter it is shown that the iron moiety of sodium nitroprusside, rather than nitric oxide inhibits mineralization. Finally, chapter 6 summarizes the main conclusions of this thesis.

Table 3.

Proteins that have been associated with mineralization[128]

Protein Mouse Mutant

Phenotype In vitro effect on mineralization References

Annexin V Anxa5 -/- No obvious altered phenotype Mediates calcium transport over a membrane, and siRNA inhibits mineralization

[72-74;77]

APase TNAP -/- Hypophosphatemia and growth impaired

Cleaves organic Pi to inorganic Pi [71;166;167]

BSP

ND* ND* Nucleates calcium Pi [51;52;168]BAG-75 ND* ND* Sequesters millimolar quantities of Pi [51;52] DMP-1 dmp1 -/- Decreased mineral to matrix ratio in

bones When nonphosphorylated facilitates hydroxyapatite crystal growth

[95;101;111;112]

Fibronectin ND* Non viable Facilitates hydroxyapatite crystal growth [113;114] Fetuin Fetuin -/- Vascular and soft tissue calcification

Inhibits precipitation of calcium and Pi

[119;121;122]

Matrix GLA protein MGP -/- Vascular, valve and cartilage calcification

ND* [123]

OPN OPN -/- Enhanced valve implants calcification

When phosphorylated inhibits mineralization

[128;129]

ON ON -/- Increased volume of adipose tissue May retard crystal growth [96;134;135;169] OC OC -/- Increased bone mass May retard crystal growth [96;134-136] * ND means not determined

Chapter 1

18

Figure 3. The ATDC5 cell culture system. A) ATDC5 cells cultured for 1 day have a fibroblast-like phenotype, and (B) develop nodules (high densities of aggregated cells with a round shape) when grown for at least 14 days. C) Spontaneous mineralization is detected when cells are grown for 45 days. D) Mineralization is detected after 14 days when extra Pi is added to the medium for the last 24 hours. Cell cultures were stained with alizarin red. The red color indicates mineralization spots. Reference list 1. Kronenberg HM (2003) Developmental regulation of the growth plate. Nature 423:332-336 2. Giachelli CM (1999) Ectopic calcification: gathering hard facts about soft tissue

mineralization. Am.J.Pathol. 154:671-675 3. Giachelli CM (2005) Inducers and inhibitors of biomineralization: lessons from

pathological calcification. Orthod.Craniofac.Res. 8:229-231 4. Terkeltaub RA (2002) What does cartilage calcification tell us about osteoarthritis?

J.Rheumatol. 29:411-415 5. Karpouzas GA, Terkeltaub RA (1999) New developments in the pathogenesis of articular

cartilage calcification. Curr.Rheumatol.Rep. 1:121-127 6. Maldonado I, Reginato AM, Reginato AJ (2001) Familial calcium crystal diseases: what

have we learned? Curr.Opin.Rheumatol. 13:225-233 7. Zhang Y, Brown MA (2005) Genetic studies of chondrocalcinosis. Curr.Opin.Rheumatol.

17:330-335 8. Wilson PW, Kauppila LI, O'Donnell CJ, Kiel DP, Hannan M, Polak JM, Cupples LA (2001)

Abdominal aortic calcific deposits are an important predictor of vascular morbidity and mortality. Circulation 103:1529-1534

9. Rutsch F, Terkeltaub R (2005) Deficiencies of physiologic calcification inhibitors and low-grade inflammation in arterial calcification: lessons for cartilage calcification. Joint Bone Spine 72:110-118


19

10. Giachelli CM (2003) Vascular calcification: in vitro evidence for the role of inorganic phosphate. J.Am.Soc.Nephrol. 14:S300-S304

11. Magne D, Julien M, Vinatier C, Merhi-Soussi F, Weiss P, Guicheux J (2005) Cartilage formation in growth plate and arteries: from physiology to pathology. Bioessays 27:708-716

12. Olsen BR, Reginato AM, Wang W (2000) Bone development. Annu.Rev.Cell Dev.Biol. 16:191-220

13. Behonick DJ, Werb Z (2003) A bit of give and take: the relationship between the extracellular matrix and the developing chondrocyte. Mech.Dev. 120:1327-1336

14. Mackie EJ (2003) Osteoblasts: novel roles in orchestration of skeletal architecture. Int.J.Biochem.Cell Biol. 35:1301-1305

15. Beck GR, Jr. (2003) Inorganic phosphate as a signaling molecule in osteoblast differentiation. J.Cell Biochem. 90:234-243

16. Franz-Odendaal TA, Hall BK, Witten PE (2006) Buried alive: how osteoblasts become osteocytes. Dev.Dyn. 235:176-190

17. Knothe Tate ML, Adamson JR, Tami AE, Bauer TW (2004) The osteocyte. Int.J.Biochem.Cell Biol. 36:1-8

18. Ducy P, Zhang R, Geoffroy V, Ridall AL, Karsenty G (1997) Osf2/Cbfa1: a transcriptional activator of osteoblast differentiation. Cell 89:747-754

19. Komori T, Yagi H, Nomura S, Yamaguchi A, Sasaki K, Deguchi K, Shimizu Y, Bronson RT, Gao YH, Inada M, Sato M, Okamoto R, Kitamura Y, Yoshiki S, Kishimoto T (1997) Targeted disruption of Cbfa1 results in a complete lack of bone formation owing to maturational arrest of osteoblasts. Cell 89:755-764

20. Liu W, Toyosawa S, Furuichi T, Kanatani N, Yoshida C, Liu Y, Himeno M, Narai S, Yamaguchi A, Komori T (2001) Overexpression of Cbfa1 in osteoblasts inhibits osteoblast maturation and causes osteopenia with multiple fractures. J.Cell Biol. 155:157-166

21. Karsenty G (2001) Minireview: transcriptional control of osteoblast differentiation. Endocrinology 142:2731-2733

22. Martin TJ, Sims NA (2005) Osteoclast-derived activity in the coupling of bone formation to resorption. Trends Mol.Med. 11:76-81

23. Shapiro IM, Adams CS, Freeman T, Srinivas V (2005) Fate of the hypertrophic chondrocyte: microenvironmental perspectives on apoptosis and survival in the epiphyseal growth plate. Birth Defects Res.C.Embryo.Today 75:330-339

24. Erlebacher A, Filvaroff EH, Gitelman SE, Derynck R (1995) Toward a molecular understanding of skeletal development. Cell 80:371-378

25. Shapiro IM, Boyde A (1984) Microdissection--elemental analysis of the mineralizing growth cartilage of the normal and rachitic chick. Metab Bone Dis.Relat Res. 5:317-326

26. Wuthier RE (1993) Involvement of cellular metabolism of calcium and phosphate in calcification of avian growth plate cartilage. J.Nutr. 123:301-309

27. White A, Wallis G (2001) Endochondral ossification: a delicate balance between growth and mineralisation. Curr.Biol. 11:R589-R591

28. Goldring MB, Tsuchimochi K, Ijiri K (2006) The control of chondrogenesis. J.Cell Biochem. 97:33-44

29. Foster JW, Dominguez-Steglich MA, Guioli S, Kowk G, Weller PA, Stevanovic M, Weissenbach J, Mansour S, Young ID, Goodfellow PN, . (1994) Campomelic dysplasia and autosomal sex reversal caused by mutations in an SRY-related gene. Nature 372:525-530

30. Wagner T, Wirth J, Meyer J, Zabel B, Held M, Zimmer J, Pasantes J, Bricarelli FD, Keutel J, Hustert E, . (1994) Autosomal sex reversal and campomelic dysplasia are caused by mutations in and around the SRY-related gene SOX9. Cell 79:1111-1120

Chapter 1

20

31. Bi W, Deng JM, Zhang Z, Behringer RR, de CB (1999) Sox9 is required for cartilage formation. Nat.Genet. 22:85-89

32. Bianco P, Cancedda FD, Riminucci M, Cancedda R (1998) Bone formation via cartilage models: the "borderline" chondrocyte. Matrix Biol. 17:185-192

33. Galotto M, Campanile G, Robino G, Cancedda FD, Bianco P, Cancedda R (1994) Hypertrophic chondrocytes undergo further differentiation to osteoblast-like cells and participate in the initial bone formation in developing chick embryo. J.Bone Miner.Res. 9:1239-1249

34. Riminucci M, Bradbeer JN, Corsi A, Gentili C, Descalzi F, Cancedda R, Bianco P (1998) Vis-a-vis cells and the priming of bone formation. J.Bone Miner.Res. 13:1852-1861

35. Thesingh CW, Groot CG, Wassenaar AM (1991) Transdifferentiation of hypertrophic chondrocytes into osteoblasts in murine fetal metatarsal bones, induced by co-cultured cerebrum. Bone Miner. 12:25-40

36. Speer MY, Giachelli CM (2004) Regulation of cardiovascular calcification. Cardiovasc.Pathol. 13:63-70

37. Demer LL, Tintut Y (2003) Mineral exploration: search for the mechanism of vascular calcification and beyond: the 2003 Jeffrey M. Hoeg Award lecture. Arterioscler.Thromb.Vasc.Biol. 23:1739-1743

38. Kirsch T (2006) Determinants of pathological mineralization. Curr.Opin.Rheumatol. 18:174-180

39. Bobryshev YV (2005) Transdifferentiation of smooth muscle cells into chondrocytes in atherosclerotic arteries in situ: implications for diffuse intimal calcification. J.Pathol. 205:641-650

40. Tyson KL, Reynolds JL, McNair R, Zhang Q, Weissberg PL, Shanahan CM (2003) Osteo/chondrocytic transcription factors and their target genes exhibit distinct patterns of expression in human arterial calcification. Arterioscler.Thromb.Vasc.Biol. 23:489-494

41. Steitz SA, Speer MY, Curinga G, Yang HY, Haynes P, Aebersold R, Schinke T, Karsenty G, Giachelli CM (2001) Smooth muscle cell phenotypic transition associated with calcification: upregulation of Cbfa1 and downregulation of smooth muscle lineage markers. Circ.Res. 89:1147-1154

42. Shioi A, Nishizawa Y, Jono S, Koyama H, Hosoi M, Morii H (1995) Beta-glycerophosphate accelerates calcification in cultured bovine vascular smooth muscle cells. Arterioscler.Thromb.Vasc.Biol. 15:2003-2009

43. Bostrom K, Watson KE, Horn S, Wortham C, Herman IM, Demer LL (1993) Bone morphogenetic protein expression in human atherosclerotic lesions. J.Clin.Invest 91:1800-1809

44. Deneke T, Langner K, Grewe PH, Harrer E, Muller KM (2001) Ossification in atherosclerotic carotid arteries. Z.Kardiol. 90 Suppl 3:106-115

45. Seifert G (1997) [Heterotopic (extraosseous) calcification (calcinosis). Etiology, pathogenesis and clinical importance]. Pathologe 18:430-438

46. Block GA, Hulbert-Shearon TE, Levin NW, Port FK (1998) Association of serum phosphorus and calcium x phosphate product with mortality risk in chronic hemodialysis patients: a national study. Am.J.Kidney Dis. 31:607-617

47. Anderson HC, Sipe JB, Hessle L, Dhanyamraju R, Atti E, Camacho NP, Millan JL (2004) Impaired calcification around matrix vesicles of growth plate and bone in alkaline phosphatase-deficient mice. Am.J.Pathol. 164:841-847

48. Anderson HC (1995) Molecular biology of matrix vesicles. Clin.Orthop.Relat Res.266-280 49. Gibson G (1998) Active role of chondrocyte apoptosis in endochondral ossification.

Microsc.Res.Tech. 43:191-204


21

50. Glimcher MJ (1989) Mechanism of calcification: role of collagen fibrils and collagen-phosphoprotein complexes in vitro and in vivo. Anat.Rec. 224:139-153

51. Gorski JP, Wang A, Lovitch D, Law D, Powell K, Midura RJ (2004) Extracellular bone acidic glycoprotein-75 defines condensed mesenchyme regions to be mineralized and localizes with bone sialoprotein during intramembranous bone formation. J.Biol.Chem. 279:25455-25463

52. Midura RJ, Wang A, Lovitch D, Law D, Powell K, Gorski JP (2004) Bone acidic glycoprotein-75 delineates the extracellular sites of future bone sialoprotein accumulation and apatite nucleation in osteoblastic cultures. J.Biol.Chem. 279:25464-25473

53. Schinke T, Karsenty G (2000) Vascular calcification--a passive process in need of inhibitors. Nephrol.Dial.Transplant. 15:1272-1274

54. Borg TK, Runyan R, Wuthier RE (1981) A freeze-fracture study of avian epiphyseal cartilage differentiation. Anat.Rec. 199:449-457

55. Hoshi K, Ozawa H (2000) Matrix vesicle calcification in bones of adult rats. Calcif.Tissue Int. 66:430-434

56. Reynolds JL, Joannides AJ, Skepper JN, McNair R, Schurgers LJ, Proudfoot D, Jahnen-Dechent W, Weissberg PL, Shanahan CM (2004) Human vascular smooth muscle cells undergo vesicle-mediated calcification in response to changes in extracellular calcium and phosphate concentrations: a potential mechanism for accelerated vascular calcification in ESRD. J.Am.Soc.Nephrol. 15:2857-2867

57. Wuthier RE, Chin JE, Hale JE, Register TC, Hale LV, Ishikawa Y (1985) Isolation and characterization of calcium-accumulating matrix vesicles from chondrocytes of chicken epiphyseal growth plate cartilage in primary culture. J.Biol.Chem. 260:15972-15979

58. Kirsch T, Wuthier RE (1994) Stimulation of calcification of growth plate cartilage matrix vesicles by binding to type II and X collagens. J.Biol.Chem. 269:11462-11469

59. Kirsch T, Nah HD, Shapiro IM, Pacifici M (1997) Regulated production of mineralization-competent matrix vesicles in hypertrophic chondrocytes. J.Cell Biol. 137:1149-1160

60. Kirsch T, Harrison G, Golub EE, Nah HD (2000) The roles of annexins and types II and X collagen in matrix vesicle-mediated mineralization of growth plate cartilage. J.Biol.Chem. 275:35577-35583

61. Kirsch T, Wang W, Pfander D (2003) Functional differences between growth plate apoptotic bodies and matrix vesicles. J.Bone Miner.Res. 18:1872-1881

62. Hsu HH, Anderson HC (1996) Evidence of the presence of a specific ATPase responsible for ATP-initiated calcification by matrix vesicles isolated from cartilage and bone. J.Biol.Chem. 271:26383-26388

63. Wu LN, Sauer GR, Genge BR, Valhmu WB, Wuthier RE (2003) Effects of analogues of inorganic phosphate and sodium ion on mineralization of matrix vesicles isolated from growth plate cartilage of normal rapidly growing chickens. J.Inorg.Biochem. 94:221-235

64. Hsu HH, Camacho NP (1999) Isolation of calcifiable vesicles from human atherosclerotic aortas. Atherosclerosis 143:353-362

65. Hsu HH, Tawfik O, Sun F (2002) Mechanisms of calcification by vesicles isolated from atherosclerotic rabbit aortas. Biochim.Biophys.Acta 1563:18-22

66. Hsu HH, Camacho NC, Tawfik O, Sun F (2002) Induction of calcification in rabbit aortas by high cholesterol diets: roles of calcifiable vesicles in dystrophic calcification. Atherosclerosis 161:85-94

67. Hsu HH, Camacho NP, Sun F, Tawfik O, Aono H (2000) Isolation of calcifiable vesicles from aortas of rabbits fed with high cholesterol diets. Atherosclerosis 153:337-348

Chapter 1

22

68. Majeska RJ, Holwerda DL, Wuthier RE (1979) Localization of phosphatidylserine in isolated chick epiphyseal cartilage matrix vesicles with trinitrobenzenesulfonate. Calcif.Tissue Int. 27:41-46

69. Palmer G, Zhao J, Bonjour J, Hofstetter W, Caverzasio J (1999) In vivo expression of transcripts encoding the Glvr-1 phosphate transporter/retrovirus receptor during bone development. Bone 24:1-7

70. Montessuit C, Bonjour JP, Caverzasio J (1995) Expression and regulation of Na-dependent P(i) transport in matrix vesicles produced by osteoblast-like cells. J.Bone Miner.Res. 10:625-631

71. Whyte MP (1994) Hypophosphatasia and the role of alkaline phosphatase in skeletal mineralization. Endocr.Rev. 15:439-461

72. Kirsch T, Nah HD, Demuth DR, Harrison G, Golub EE, Adams SL, Pacifici M (1997) Annexin V-mediated calcium flux across membranes is dependent on the lipid composition: implications for cartilage mineralization. Biochemistry 36:3359-3367

73. Matsuda R, Kaneko N, Horikawa Y (1997) Presence and comparison of Ca2+ transport activity of annexins I, II, V, and VI in large unilamellar vesicles. Biochem.Biophys.Res.Commun. 237:499-503

74. Wang W, Xu J, Kirsch T (2005) Annexin V and terminal differentiation of growth plate chondrocytes. Exp.Cell Res. 305:156-165

75. Balcerzak M, Hamade E, Zhang L, Pikula S, Azzar G, Radisson J, Bandorowicz-Pikula J, Buchet R (2003) The roles of annexins and alkaline phosphatase in mineralization process. Acta Biochim.Pol. 50:1019-1038

76. Kirsch T, Swoboda B, Nah H (2000) Activation of annexin II and V expression, terminal differentiation, mineralization and apoptosis in human osteoarthritic cartilage. Osteoarthritis.Cartilage. 8:294-302

77. Brachvogel B, Dikschas J, Moch H, Welzel H, von der MK, Hofmann C, Poschl E (2003) Annexin A5 is not essential for skeletal development. Mol.Cell Biol. 23:2907-2913

78. Magne D, Bluteau G, Faucheux C, Palmer G, Vignes-Colombeix C, Pilet P, Rouillon T, Caverzasio J, Weiss P, Daculsi G, Guicheux J (2003) Phosphate is a specific signal for ATDC5 chondrocyte maturation and apoptosis-associated mineralization: possible implication of apoptosis in the regulation of endochondral ossification. J.Bone Miner.Res. 18:1430-1442

79. Mansfield K, Rajpurohit R, Shapiro IM (1999) Extracellular phosphate ions cause apoptosis of terminally differentiated epiphyseal chondrocytes. J.Cell Physiol 179:276-286

80. Mansfield K, Teixeira CC, Adams CS, Shapiro IM (2001) Phosphate ions mediate chondrocyte apoptosis through a plasma membrane transporter mechanism. Bone 28:1-8

81. Mansfield K, Pucci B, Adams CS, Shapiro IM (2003) Induction of apoptosis in skeletal tissues: phosphate-mediated chick chondrocyte apoptosis is calcium dependent. Calcif.Tissue Int. 73:161-172

82. Sabbagh Y, Carpenter TO, Demay MB (2005) Hypophosphatemia leads to rickets by impairing caspase-mediated apoptosis of hypertrophic chondrocytes. Proc.Natl.Acad.Sci.U.S.A 102:9637-9642

83. Ketteler M, Brandenburg V, Jahnen-Dechent W, Westenfeld R, Floege J (2005) Do not be misguided by guidelines: the calcium x phosphate product can be a Trojan horse. Nephrol.Dial.Transplant. 20:673-677

84. Takeda E, Yamamoto H, Nashiki K, Sato T, Arai H, Taketani Y (2004) Inorganic phosphate homeostasis and the role of dietary phosphorus. J.Cell Mol.Med. 8:191-200


23

85. Lomashvili KA, Cobbs S, Hennigar RA, Hardcastle KI, O'Neill WC (2004) Phosphate-induced vascular calcification: role of pyrophosphate and osteopontin. J.Am.Soc.Nephrol. 15:1392-1401

86. Bratton DL, Fadok VA, Richter DA, Kailey JM, Guthrie LA, Henson PM (1997) Appearance of phosphatidylserine on apoptotic cells requires calcium-mediated nonspecific flip-flop and is enhanced by loss of the aminophospholipid translocase. J.Biol.Chem. 272:26159-26165

87. Schoen FJ, Tsao JW, Levy RJ (1986) Calcification of bovine pericardium used in cardiac valve bioprostheses. Implications for the mechanisms of bioprosthetic tissue mineralization. Am.J.Pathol. 123:134-145

88. Boskey AL, Dick BL (1991) The effect of phosphatidylserine on in vitro hydroxyapatite growth and proliferation. Calcif.Tissue Int. 49:193-196

89. Cheung HS, Ryan LM (1999) Phosphocitrate blocks nitric oxide-induced calcification of cartilage and chondrocyte-derived apoptotic bodies. Osteoarthritis.Cartilage. 7:409-412

90. Hashimoto S, Ochs RL, Rosen F, Quach J, McCabe G, Solan J, Seegmiller JE, Terkeltaub R, Lotz M (1998) Chondrocyte-derived apoptotic bodies and calcification of articular cartilage. Proc.Natl.Acad.Sci.U.S.A 95:3094-3099

91. Lotz M, Hashimoto S, Kuhn K (1999) Mechanisms of chondrocyte apoptosis. Osteoarthritis.Cartilage. 7:389-391

92. Proudfoot D, Skepper JN, Hegyi L, Bennett MR, Shanahan CM, Weissberg PL (2000) Apoptosis regulates human vascular calcification in vitro: evidence for initiation of vascular calcification by apoptotic bodies. Circ.Res. 87:1055-1062

93. Roach HI, Aigner T, Kouri JB (2004) Chondroptosis: a variant of apoptotic cell death in chondrocytes? Apoptosis. 9:265-277

94. Roach HI, Clarke NM (2000) Physiological cell death of chondrocytes in vivo is not confined to apoptosis. New observations on the mammalian growth plate. J.Bone Joint Surg.Br. 82:601-613

95. He G, Gajjeraman S, Schultz D, Cookson D, Qin C, Butler WT, Hao J, George A (2005) Spatially and temporally controlled biomineralization is facilitated by interaction between self-assembled dentin matrix protein 1 and calcium phosphate nuclei in solution. Biochemistry 44:16140-16148

96. Hunter GK, Hauschka PV, Poole AR, Rosenberg LC, Goldberg HA (1996) Nucleation and inhibition of hydroxyapatite formation by mineralized tissue proteins. Biochem.J. 317 ( Pt 1):59-64

97. Hunter GK, Poitras MS, Underhill TM, Grynpas MD, Goldberg HA (2001) Induction of collagen mineralization by a bone sialoprotein--decorin chimeric protein. J.Biomed.Mater.Res. 55:496-502

98. Butler WT, Finch JE, Jr., Desteno CV (1972) Chemical character of proteins in rat incisors. Biochim.Biophys.Acta 257:167-171

99. Donley GE, Fitzpatrick LA (1998) Noncollagenous matrix proteins controlling mineralization; possible role in pathologic calcification of vascular tissue. Trends Cardiovasc.Med. 8:199-206

100. He G, Dahl T, Veis A, George A (2003) Nucleation of apatite crystals in vitro by self-assembled dentin matrix protein 1. Nat.Mater. 2:552-558

101. Tartaix PH, Doulaverakis M, George A, Fisher LW, Butler WT, Qin C, Salih E, Tan M, Fujimoto Y, Spevak L, Boskey AL (2004) In vitro effects of dentin matrix protein-1 on hydroxyapatite formation provide insights into in vivo functions. J.Biol.Chem. 279:18115-18120

Chapter 1

24

102. Bianco P, Fisher LW, Young MF, Termine JD, Robey PG (1991) Expression of bone sialoprotein (BSP) in developing human tissues. Calcif.Tissue Int. 49:421-426

103. Chen JK, Shapiro HS, Wrana JL, Reimers S, Heersche JN, Sodek J (1991) Localization of bone sialoprotein (BSP) expression to sites of mineralized tissue formation in fetal rat tissues by in situ hybridization. Matrix 11:133-143

104. McQuillan DJ, Richardson MD, Bateman JF (1995) Matrix deposition by a calcifying human osteogenic sarcoma cell line (SAOS-2). Bone 16:415-426

105. Stanford CM, Jacobson PA, Eanes ED, Lembke LA, Midura RJ (1995) Rapidly forming apatitic mineral in an osteoblastic cell line (UMR 106-01 BSP). J.Biol.Chem. 270:9420-9428

106. Wang A, Martin JA, Lembke LA, Midura RJ (2000) Reversible suppression of in vitro biomineralization by activation of protein kinase A. J.Biol.Chem. 275:11082-11091

107. Gorski JP, Kremer EA, Chen Y, Ryan S, Fullenkamp C, Delviscio J, Jensen K, McKee MD (1997) Bone acidic glycoprotein-75 self-associates to form macromolecular complexes in vitro and in vivo with the potential to sequester phosphate ions. J.Cell Biochem. 64:547-564

108. Wang D, Christensen K, Chawla K, Xiao G, Krebsbach PH, Franceschi RT (1999) Isolation and characterization of MC3T3-E1 preosteoblast subclones with distinct in vitro and in vivo differentiation/mineralization potential. J.Bone Miner.Res. 14:893-903

109. D'Souza RN, Cavender A, Sunavala G, Alvarez J, Ohshima T, Kulkarni AB, MacDougall M (1997) Gene expression patterns of murine dentin matrix protein 1 (Dmp1) and dentin sialophosphoprotein (DSPP) suggest distinct developmental functions in vivo. J.Bone Miner.Res. 12:2040-2049

110. George A, Silberstein R, Veis A (1995) In situ hybridization shows Dmp1 (AG1) to be a developmentally regulated dentin-specific protein produced by mature odontoblasts. Connect.Tissue Res. 33:67-72

111. MacDougall M, Gu TT, Luan X, Simmons D, Chen J (1998) Identification of a novel isoform of mouse dentin matrix protein 1: spatial expression in mineralized tissues. J.Bone Miner.Res. 13:422-431

112. Ling Y, Rios HF, Myers ER, Lu Y, Feng JQ, Boskey AL (2005) DMP1 depletion decreases bone mineralization in vivo: an FTIR imaging analysis. J.Bone Miner.Res. 20:2169-2177

113. Daculsi G, Pilet P, Cottrel M, Guicheux G (1999) Role of fibronectin during biological apatite crystal nucleation: ultrastructural characterization. J.Biomed.Mater.Res. 47:228-233

114. Couchourel D, Escoffier C, Rohanizadeh R, Bohic S, Daculsi G, Fortun Y, Padrines M (1999) Effects of fibronectin on hydroxyapatite formation. J.Inorg.Biochem. 73:129-136

115. Ketteler M, Vermeer C, Wanner C, Westenfeld R, Jahnen-Dechent W, Floege J (2002) Novel insights into uremic vascular calcification: role of matrix Gla protein and alpha-2-Heremans Schmid glycoprotein/fetuin. Blood Purif. 20:473-476

116. Brown WM, Saunders NR, Mollgard K, Dziegielewska KM (1992) Fetuin--an old friend revisited. Bioessays 14:749-755

117. Wendel M, Heinegard D, Franzen A (1993) A major non-collagenous 62 kDa protein from rat bone mineralized matrix is identical to pp63 a phosphorylated glycoprotein from liver. Matrix 13:331-339

118. Takagi Y, Shimokawa H, Suzuki M, Nagai H, Sasaki S (1990) Immunohistochemical localization of alpha 2HS glycoprotein in dentin. Calcif.Tissue Int. 47:40-45

119. Merx MW, Schafer C, Westenfeld R, Brandenburg V, Hidajat S, Weber C, Ketteler M, Jahnen-Dechent W (2005) Myocardial stiffness, cardiac remodeling, and diastolic dysfunction in calcification-prone fetuin-A-deficient mice. J.Am.Soc.Nephrol. 16:3357-3364


25

120. Ketteler M (2005) Fetuin-A and extraosseous calcification in uremia. Curr.Opin.Nephrol.Hypertens. 14:337-342

121. Schinke T, Amendt C, Trindl A, Poschke O, Muller-Esterl W, Jahnen-Dechent W (1996) The serum protein alpha2-HS glycoprotein/fetuin inhibits apatite formation in vitro and in mineralizing calvaria cells. A possible role in mineralization and calcium homeostasis. J.Biol.Chem. 271:20789-20796

122. Price PA, Lim JE (2003) The inhibition of calcium phosphate precipitation by fetuin is accompanied by the formation of a fetuin-mineral complex. J.Biol.Chem. 278:22144-22152

123. Luo G, Ducy P, McKee MD, Pinero GJ, Loyer E, Behringer RR, Karsenty G (1997) Spontaneous calcification of arteries and cartilage in mice lacking matrix GLA protein. Nature 386:78-81

124. Reynolds JL, Skepper JN, McNair R, Kasama T, Gupta K, Weissberg PL, Jahnen-Dechent W, Shanahan CM (2005) Multifunctional roles for serum protein fetuin-a in inhibition of human vascular smooth muscle cell calcification. J.Am.Soc.Nephrol. 16:2920-2930

125. Boskey AL, Maresca M, Ullrich W, Doty SB, Butler WT, Prince CW (1993) Osteopontin-hydroxyapatite interactions in vitro: inhibition of hydroxyapatite formation and growth in a gelatin-gel. Bone Miner. 22:147-159

126. Beck GR, Jr., Moran E, Knecht N (2003) Inorganic phosphate regulates multiple genes during osteoblast differentiation, including Nrf2. Exp.Cell Res. 288:288-300

127. Steitz SA, Speer MY, McKee MD, Liaw L, Almeida M, Yang H, Giachelli CM (2002) Osteopontin inhibits mineral deposition and promotes regression of ectopic calcification. Am.J.Pathol. 161:2035-2046

128. Gericke A, Qin C, Spevak L, Fujimoto Y, Butler WT, Sorensen ES, Boskey AL (2005) Importance of phosphorylation for osteopontin regulation of biomineralization. Calcif.Tissue Int. 77:45-54

129. Giachelli CM, Steitz S (2000) Osteopontin: a versatile regulator of inflammation and biomineralization. Matrix Biol. 19:615-622

130. Jono S, Peinado C, Giachelli CM (2000) Phosphorylation of osteopontin is required for inhibition of vascular smooth muscle cell calcification. J.Biol.Chem. 275:20197-20203

131. Sodek J, Ganss B, McKee MD (2000) Osteopontin. Crit Rev.Oral Biol.Med. 11:279-303

132. Alford AI, Hankenson KD (2006) Matricellular proteins: Extracellular modulators of bone development, remodeling, and regeneration. Bone

133. Doi Y, Horiguchi T, Kim SH, Moriwaki Y, Wakamatsu N, Adachi M, Ibaraki K, Moriyama K, Sasaki S, Shimokawa H (1992) Effects of non-collagenous proteins on the formation of apatite in calcium beta-glycerophosphate solutions. Arch.Oral Biol. 37:15-21

134. Menanteau J, Neuman WF, Neuman MW (1982) A study of bone proteins which can prevent hydroxyapatite formation. Metab Bone Dis.Relat Res. 4:157-162

135. Romberg RW, Werness PG, Riggs BL, Mann KG (1986) Inhibition of hydroxyapatite crystal growth by bone-specific and other calcium-binding proteins. Biochemistry 25:1176-1180

136. Ducy P, Desbois C, Boyce B, Pinero G, Story B, Dunstan C, Smith E, Bonadio J, Goldstein S, Gundberg C, Bradley A, Karsenty G (1996) Increased bone formation in osteocalcin-deficient mice. Nature 382:448-452

137. Atsumi T, Miwa Y, Kimata K, Ikawa Y (1990) A chondrogenic cell line derived from a differentiating culture of AT805 teratocarcinoma cells. Cell Differ.Dev. 30:109-116

138. Shukunami C, Shigeno C, Atsumi T, Ishizeki K, Suzuki F, Hiraki Y (1996) Chondrogenic differentiation of clonal mouse embryonic cell line ATDC5 in vitro: differentiation-

Chapter 1

26

dependent gene expression of parathyroid hormone (PTH)/PTH-related peptide receptor. J.Cell Biol. 133:457-468

139. Enomoto H, Enomoto-Iwamoto M, Iwamoto M, Nomura S, Himeno M, Kitamura Y, Kishimoto T, Komori T (2000) Cbfa1 is a positive regulatory factor in chondrocyte maturation. J.Biol.Chem. 275:8695-8702

140. Shukunami C, Ishizeki K, Atsumi T, Ohta Y, Suzuki F, Hiraki Y (1997) Cellular hypertrophy and calcification of embryonal carcinoma-derived chondrogenic cell line ATDC5 in vitro. J.Bone Miner.Res. 12:1174-1188

141. Guicheux J, Palmer G, Shukunami C, Hiraki Y, Bonjour JP, Caverzasio J (2000) A novel in vitro culture system for analysis of functional role of phosphate transport in endochondral ossification. Bone 27:69-74

142. Palmer G, Guicheux J, Bonjour JP, Caverzasio J (2000) Transforming growth factor-beta stimulates inorganic phosphate transport and expression of the type III phosphate transporter Glvr-1 in chondrogenic ATDC5 cells. Endocrinology 141:2236-2243

143. Fujita T, Meguro T, Izumo N, Yasutomi C, Fukuyama R, Nakamuta H, Koida M (2001) Phosphate stimulates differentiation and mineralization of the chondroprogenitor clone ATDC5. Jpn.J.Pharmacol. 85:278-281

144. Jang D, Murrell GA (1998) Nitric oxide in arthritis. Free Radic.Biol.Med. 24:1511-1519 145. van't Hof RJ, Ralston SH (2001) Nitric oxide and bone. Immunology 103:255-261 146. Archer RS, Bayley JI, Archer CW, Ali SY (1993) Cell and matrix changes associated with

pathological calcification of the human rotator cuff tendons. J.Anat. 182 ( Pt 1):1-11 147. Uhthoff HK (1996) Calcifying tendinitis. Ann.Chir Gynaecol. 85:111-115 148. Zhang Y, Brown MA (2005) Genetic studies of chondrocalcinosis. Curr.Opin.Rheumatol.

17:330-335 149. Crevenna R, Keilani M, Wiesinger G, Nicolakis P, Quittan M, Fialka-Moser V (2002)

Calcific trochanteric bursitis: resolution of calcifications and clinical remission with non-invasive treatment. A case report. Wien.Klin.Wochenschr. 114:345-348

150. Rufai A, Ralphs JR, Benjamin M (1995) Structure and histopathology of the insertional region of the human Achilles tendon. J.Orthop.Res. 13:585-593

151. Stahnke M, Mangham DC, Davies AM (2004) Calcific haemorrhagic bursitis anterior to the knee mimicking a soft tissue sarcoma: report of two cases. Skeletal Radiol. 33:363-366

152. Peng B, Hou S, Shi Q, Jia L (2000) Experimental study on mechanism of vertebral osteophyte formation. Chin J.Traumatol. 3:202-205

153. Thurston AJ (2002) Bone spurs: mechanism of production of different shapes based on observations in Dupuytren's diathesis. ANZ.J.Surg. 72:290-293

154. Mohler ER, III (2004) Mechanisms of aortic valve calcification. Am.J.Cardiol. 94:1396-402, A6

155. Vattikuti R, Towler DA (2004) Osteogenic regulation of vascular calcification: an early perspective. Am.J.Physiol Endocrinol.Metab 286:E686-E696

156. Vanden BL, Vanderstraeten G (2005) Heterotopic ossification: a review. J.Rehabil.Med. 37:129-136

157. Chan ED, Morales DV, Welsh CH, McDermott MT, Schwarz MI (2002) Calcium deposition with or without bone formation in the lung. Am.J.Respir.Crit Care Med. 165:1654-1669

158. Chung MJ, Lee KS, Franquet T, Muller NL, Han J, Kwon OJ (2005) Metabolic lung disease: imaging and histopathologic findings. Eur.J Radiol. 54:233-245

159. Conlin PA, Jimenez-Quintero LP, Rapini RP (2002) Osteomas of the skin revisited: a clinicopathologic review of 74 cases. Am.J.Dermatopathol. 24:479-483


27

160. Chen KH, Cheng WT, Li MJ, Yang DM, Lin SY (2005) Calcification of senile cataractous lens determined by Fourier transform infrared (FTIR) and Raman microspectroscopies. J.Microsc. 219:36-41

161. Moe OW (2006) Kidney stones: pathophysiology and medical management. Lancet 367:333-344

162. Baba Y, Broderick DF, Uitri RJ, Hutton ML, Wszolek ZK (2005) Heredofamilial brain calcinosis syndrome. Mayo Clin.Proc. 80:641-651

163. Manyam BV (2005) What is and what is not 'Fahr's disease'. Parkinsonism.Relat Disord. 11:73-80

164. van Kuijk AA, Geurts AC, van Kuppevelt HJ (2002) Neurogenic heterotopic ossification in spinal cord injury. Spinal Cord. 40:313-326

165. Wang LL (2005) Biology of osteogenic sarcoma. Cancer J. 11:294-305 166. Narisawa S, Frohlander N, Millan JL (1997) Inactivation of two mouse alkaline

phosphatase genes and establishment of a model of infantile hypophosphatasia. Dev.Dyn. 208:432-446

167. Tesch W, Vandenbos T, Roschgr P, Fratzl-Zelman N, Klaushofer K, Beertsen W, Fratzl P (2003) Orientation of mineral crystallites and mineral density during skeletal development in mice deficient in tissue nonspecific alkaline phosphatase. J.Bone Miner.Res. 18:117-125

168. Hunter GK, Goldberg HA (1993) Nucleation of hydroxyapatite by bone sialoprotein. Proc.Natl.Acad.Sci.U.S.A 90:8562-8565

169. Framson PE, Sage EH (2004) SPARC and tumor growth: where the seed meets the soil? J.Cell Biochem. 92:679-690

Chapter 2

Rapid cell-mediated large mineralizing structures in a serum free cell culture system.

Leonie F.A. Huitema1,2, P. René van Weeren2, T. Visser3, Chris H.A. van de Lest1,2, Ab Barneveld2, J. Bernd Helms1, Arie B. Vaandrager1

1 Department of Biochemistry and Cell Biology, 2 Department of Equine Sciences, Faculty of Veterinary Medicine, and Graduate School Animal Health, 3 Department of inorganic chemistry and catalyses, Debye Institute, Utrecht University, Utrecht,

the Netherlands

Submitted

Chapter 2

30

Abstract The aim of this study is to investigate how cell-mediated mineralization is initiated in a cell culture system. Therefore, we used a cell culture model wherein mineralizing structures could be separated from the extracellular matrix. Subconfluent cultured ATDC5 cells were incubated for 2 hours in medium containing additional phosphate without the presence of serum, which is known to contain mineralization inhibitors. The experimental culture conditions resulted in the formation of large (sub millimetre size range) mineralizing structures (LMS) in the medium. LMS were shown to contain whole cells, which were embedded in hydroxyapatite and observed to have a stretched morphology. Staurosporin- and sodiumazide treated cells were not able to generate LMS, which suggests that cells need to be alive in order to form LMS. The amount of DNA integrated in LMS was 0.5% of the total cell population, suggesting that only a small fraction of cells is competent to generate LMS and integrate into these structures. The phosphate transporter inhibitor phosphonoformic acid was able to inhibit the development of LMS. Annexin V was translocated from the cytosol to the membrane after addition of phosphate, indicating that phosphate affects processes inside the cells. In the presence of serum no LMS were formed, but multiple small mineralized structures were observed inside the cell. Taken together, the results of this study show that excluding serum in the cell culture system enhances rapid crystal growth and suggest that cell-mediated mineralization may start intracellularly.

Cell-mediated large mineralizing structures

31

Introduction In organisms, mineralization is an essential requirement for the development of hard tissues, such as calcified cartilage, bone and dentine [1-3]. Mineralized tissues in organisms are biocomposites of a structured organic matrix impregnated with matrix oriented apatite crystals [2;3]. The organism controls the nature, orientation, size and shape of the mineral phase by creating closed compartments or defined channels in which the mineral crystals form [4]. Soft tissues do not mineralize under normal conditions [5;6]. However, under certain pathological conditions some organs, in particular blood vessels, are prone to mineralization as well [7]. For example, atherosclerosis, diabetes and end-stage renal disease predispose arteries to calcification, which is highly associated with mortality in patients with cardiovascular disease [5]. Growing evidence suggests that vascular calcification, like bone formation, is a highly regulated process, involving both inducers and inhibitors [7-9]. For example, fetuin-A is the major glycoprotein in serum known to systemically prevent unwanted calcification [10;11]. Fetuin-A knockout mice were shown to develop severe calcification of various organs [12;13]. There are two not mutually exclusive theories about the induction of mineralization. One theory proposes that mineralization is initiated within the ‘hole zone’ of collagen fibrils requiring specific non-collagenous proteins (NCPs). The NCPs are reported to constitute 5-10% of the total extracellular matrix and some of them have been implicated in the regulation of mineral deposition in dentinogenesis or osteogenesis [3]. In vitro studies have shown that non-collagenous phosphoproteins, like bone sialoprotein (BSP) or bone acidic glycoprotein-75, can nucleate the formation of apatite crystals [14-16]. However, other phosphoproteins, like osteopontin, have been shown to be inhibitors of mineralization [6;17]. NCPs have also been found in calcifying atherosclerotic plaques [18;19]. Thus, in this theory it is hypothesized that mineralization is dependent on the quality of the extracellular matrix. Another hypothesis proposes that mineralization is induced within matrix vesicles (MVs) [20]. MVs are cell-derived extracellular membrane enclosed particles, about 0.1- 1 µm in size, that initiate mineralization by the action of MVs-associated phosphatases and calcium-binding phospholipids and proteins [20]. Matrix vesicles have been isolated from cultures of mineralizing osteoblasts, chondrocytes and vascular smooth muscle cells [21-23]. Annexin V has been reported to play a major role in the function of the vesicles, because it mediates the influx of calcium inside the vesicles, which in turn permits mineral growth from a preexisting nucleation core complex [24-26]. The aim of this study was to investigate how mineralization is initiated in a cell culture system using the ATDC5 cell line [27]. It was hypothesized that serum, which is known to contain mineralization inhibitors such as fetuin, inhibits the

Chapter 2

32

early stages of mineralization in the cell culture system. Therefore, it was tested whether excluding serum from the medium shortened the cell culture period to induce mineralization, making this model system more suitable for investigation. As mineralizing structures are strongly integrated in the extracellular matrix making it difficult to investigate them in isolation [15;20], we used a cell culture system wherein mineralizing structures could be separated from the extracellular matrix. Therefore, subconfluent cultured ATDC5 cells were incubated for 2 hours in medium containing additional phosphate (Pi) without the presence of serum. This resulted in the development of large mineralizing structures (LMS) in the medium. These structures were shown to contain DNA and membranes, which seemed affect the shape of the crystal, indicating a cell-mediated effect. Materials and methods Materials Cell culture plastics were purchased from Greiner bio-one (Frickenhausen, Germany) and Nunc Inc (Rochester, NY). Fetal bovine serum (FBS), antibiotics, and Insulin-Transferrin-Selenium-X (ITS-X) were obtained from Gibco BRL (Grand Island, NY). FM 4-64 was from Invitrogen (Breda, The Netherlands), α-minimum essential medium (α-MEM) was from Biochrom AG (Berlin, Germany), Annexin-V-Alexa 568, staurosporin and western blocking reagent were from Roche Molecular Biochemicals (Mannheim, Germany). Rabbit polyclonal annexin V antibody was provided by Santa Cruz Biotechnologies Inc (Santa Cruz, CA). The Supersignal chemiluminescent substrate kit (ECL) for detection of proteins on immunoblots and BCATM Protein Assay Kit were supplied by Pierce (Cheshire, UK). All other chemicals were obtained from Sigma (Poole, UK). Cell culture ATDC5 cells (Riken Cell Bank, Tsukuba, Japan) were cultured and maintained in a 1:1 mixture of Dulbecco’s modified Eagle medium and Ham’s F12 medium (DMEM/F12) containing 5% FBS, 10 µg/ml human transferrin, 30 nM sodium selenite, 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2 – 95% air. Cells were subcultured once a week using trypsin/EDTA solution and with an initial density of 15,000 cells/cm2. For induction of differentiation, ATDC5 cells were cultured in the same medium additionally containing 10 µg/ml insulin from ITS-X (Sigma). Medium was replaced every 2 days.


33

Induction of mineralization in the absence of serum and isolation of large mineralizing structures. Subconfluent ATDC5 cells (grown for approximately 4 days in DMEM/F12 containing 5% FBS, 100 IU/ml penicillin, 100 µg/ml streptomycin and ITS-X) were washed 3 times with PBS and incubated for 2 hours in α-MEM (25 ml/ 175 cm2 flask) with or without 5 mM Pi, containing 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37°C 5% CO2. LMS were isolated from the medium by centrifuging 20 minutes at 20,000 x g in a sorvall centrifuge® RC 5B plus. Pellets were washed either in 1ml PBS (for microscopy) or Millipore H2O (for biochemical analyses) and centrifuged in a cooled eppendorf centrifuge (5402) maximum speed (15,800 x g). Induction of mineralization in the presence of serum ATDC5 cells were grown for 14 days on coverslips (ø 25 mm) in 6 well plates (ø 35 mm/well). Subsequently, mineralization was induced by transferring the cells to α-MEM containing 5% FBS, ITS-X, 5 mM phosphate, 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2 – 95% air for 2 days. Measurement of total calcium content The total amount of calcium was measured by atomic absorption spectroscopy at 422.7 nm (Perkin Elmer 3300, Perkin Elmer Corp., Norwalk, Cl, USA). Samples were washed twice with Millipore H2O, hydrolyzed in 400 µl 6 M HCL and diluted twice with 0.5% lanthanum citrate solution to control interferences with phosphate. Infrared spectroscopy Samples were washed 3 times in Millipore H2O and overnight dehydrated in a vacuum centrifuge (speedvac concentrator SVC 100H, Savant). Subsequently, samples were prepared as KBr pellets for infrared (IR) spectroscopic analysis. Spectra were recorded on a Perkin-Elmer 2000 Fourier transform infrared (FT-IR) spectrometer by accumulating 16 scans over the wavenumber range of 4000-400 cm-1 with 4 cm-1 optical resolution. For comparison, hydroxyapatite (Bio-Rad laboratories, Richmond) was also analyzed. Measurement of total DNA Total DNA was determined using the fluorescent dye Hoechst 33342 as described by Kim et al. (1988) [28]. Briefly: 200 µl 10 mM Tris, 100 mM NaCl, 10 mM EDTA (pH 7.4) containing 1 mg/L Hoechst was added to 25 µl samples which were dissolved in 60 mM EDTA, PBS (pH 7.4). The absorbance was measured at

Chapter 2

34

λex 365 nm and λem 460 nm using a microplate reader, LS 50B luminescence spectrometer (Perkin-Elmer). Annexin V translocation Cells were washed twice with cold PBS and scraped off the plate in 500 µl PBS and lysed with an ultrasone disintegrator (soniprep 50). Membrane and cytosolic fractions were separated by centrifugating the samples 1 hour at 100,000 x g in a OptimaTM Max-E Ultracentrifuge (Beckman CoulterTM, U.S.A.). Membranes were resuspended in 500 µl PBS/60 mM EDTA, and 60 mM EDTA was also added to the cytosolic fraction. Protein content was determined using the BCATM protein assay reagent kit. Proteins (20 µg) were separated by SDS-PAGE and transferred to nitrocellulose membranes. The membranes were blocked with 1% western blocking reagent overnight and subsequently 1 hour exposed to rabbit polyclonal anti-annexin V, dilution 1:500. Four washing steps with TBS-Tween (50 mM Tris; 150 mM NaCl; 0.1% Tween20, pH 7.5) followed. Blots were then incubated with horseradish peroxidase-conjugated secondary antibody for 1 hour. Subsequently, the blots were washed four times with TBS-Tween and annexin V proteins were displayed by a reaction on Supersignal chemiluminescent substrate and exposure to x-ray films. Fluorescence- and multiphoton microscopy Cells were incubated for 30 minutes with the cell membrane permeable fluorescent DNA label, Hoechst 33342 (2 µg/ml) in DMEM/F12 containing 5% FBS, 10 µg/ml human transferrin, 30 nM sodium selenite, 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37ºC 5% CO2. After incubation, cells were washed 3 times with PBS and incubated with α-MEM containing 5 mM Pi for 2 hours. Mineral crystal structures were detected by incubating the cells or LMS with 5 µg/ml calcein for at least 1 hour and subsequently washed 3 times with PBS [29;30]. To visualize membrane structures, cells or LMS were finally incubated for at least 2 minutes with FM 4-64 (5 µg/ml) according to the manufacturer’s protocol. Plasma-membrane phosphatidylserine (PS) externalization was detected by labeling cells with Annexin-V-Alexa 568 according to the manufacturer’s protocol. Isolated precipitates were imaged using a digital imaging microscope, Leica DMRA fluorescence microscope with additional photomatrix coolsnap CCD camera (Leica Microsystems BV, Rijswijk, The Netherlands). Imaging of Hoechst 33342 was visualized by λex 360/40 nm, λem 460/50, calcein at λex 460/50 nm, λem 535/50 and FM4-64 at λex 565/30 nm, λem 620/60. Coverslips containing non-fixed (living) ATDC5 cells were visualized in a 37°C incubation chamber (solent scientific, Portmouth, UK) under a Radiance 2100 MP


35

multiphoton microscope (Biorad, Veenendaal, The Netherlands) [31]. Annexin-V-Alexa 568 was imaged using a He/Ne laser laser (λex 543 nm, λem 600/50); Calcein fluorescence was imaged using an Ar laser (λex 488 nm, λem 515\30); Hoechst 33342 fluorescence was detected using a Tsunami multiphoton laser (λex 780 nm, λem 450/80). Laser intensities were kept as low as possible to prevent possible irradiation damage to the cells. Statistical analysis Experiments were performed at least three times in duplo. Statistical analyses were performed using an analysis of variance (ANOVA), followed by Dunnett's Multiple Comparison Test (one-way ANOVA). Values of P<0.05 where considered to be significant. Values where expressed as mean ± SEM.

Figure 1. Phosphate induces formation of large mineralizing structures in medium with ATDC5 cells in the absence of serum. Subconfluent ATDC5 cells were incubated in α-MEM + 5 mM Pi, or α-MEM for 2 hours. Subsequently, the medium was centrifuged for 20 minutes at 20,000 x g and the pellet was washed and analysed by phase contrast microscopy. A) Isolated LMS from medium containing 5mM Pi, B) Isolated precipitates from medium without additional Pi, notice the absence of LMS, although occasionally some death cells were observed. A representative image of 3 independent experiments is shown. C) Time course of the formation of large mineralizing structures. Precipitates were isolated from α-MEM + 5 mM Pi that was incubated for the indicated time points with subconfluent ATDC5 cells (■), or without the presence of cells (▲). Calcium was extracted and measured by atomic absorption spectrometry as described in the Materials and Methods section. The results are expressed as µg calcium/cell culture flask and represent the mean ± S.E.M. of three independent experiments

Chapter 2

36

Figure 2. Infrared spectroscopic analysis of large mineralizing structures. A) Hydroxyapatite and B) LMS obtained from subconfluent ATDC5 cells incubated for 2 hours in α-MEM + 5 mM Pi, were analyzed as KBr discs . The absorption doublets at the 1091-1029 cm-1 and 602-563 cm-1 are characteristic of HAP mineral. Results shown are representative for three independent experiments. Results Rapid formation of large mineralizing structures by addition of phosphate Serum contains fetuin, which is known to prevent mineralization [11]. Therefore, mineralization was induced by addition of 5 mM Pi in the cell culture system without the addition of serum. Furthermore, to increase the probability that mineralization structures were released into the medium, the medium switch was performed when the cells were subconfluent. This resulted in the development of large mineralizing structures (LMS) in 2 hours (Fig. 1A). LMS were estimated to be approximately in the sub millimeter range in size and were not observed when cells were incubated for 2 hours without additional Pi, although a rare occasion of a


37

death cell was detected (Fig. 1B). LMS were also neither detected in medium with 5 mM Pi without the presence of cells (data not shown), nor calcium phosphate was precipitated (Fig. 1C). Subsequently, it was investigated at what time period mineralizing structures developed by addition of 5 mM Pi in serum free medium. As shown in Fig. 1C, calcium containing crystals could be detected in the medium when incubating subconfluent ATDC5 cells at least 90 minutes in medium with 5 mM Pi. Analyses of large mineralizing structures Next, it was determined whether the LMS contained hydroxyapatite (HAP). Therefore, LMS were analyzed by FT-IR and the spectra obtained were compared to the IR spectrum of HAP. Fig. 2A shows the spectrum of HAP with two for HAP characteristic phosphate absorption doublets at 1091-1029 cm-1 and 602-563 cm-1. These doublets were also detected in the LMS spectra (Fig. 2B), indicating that LMS are HAP like structures. Next to the phosphate bands, the FT-IR spectra of LMS also showed characteristic amide peaks at 1652, 1568 and 1423 cm-1, which point to the presence of proteins in the structures. To confirm the presence of proteins, LMS were dissolved in 60 mM EDTA, separated by SDS PAGE and stained with Coomassie brilliant blue [32]. As shown in Fig. 3A, no proteins were precipitated when subconfluent ATDC5 cells were incubated in medium without additional Pi. However, proteins were detected when ATDC5 cells were incubated for at least 90 minutes in medium with 5 mM Pi. Of note, proteins were only visible on gel when LMS were dissolved in 60 mM EDTA (pH 7.4), suggesting that proteins are embedded in a HAP crystal structure in LMS. The protein bands from LMS contained a similar pattern as the protein bands from total cell lysate (Fig. 3A). Based on this observation, it was investigated whether LMS were containing cells by quantifying DNA content in LMS. Fig. 3B shows that indeed the amount of DNA increased parallel to the increase in calcium content when ATDC5 cells were incubated in medium with 5 mM Pi. The amount of DNA present in LMS was calculated to be 0.5% of the total DNA present in the cell culture flask. These results indicated that a small fraction of cells is integrated in LMS. Next, the presence of cells in LMS was visualized and nuclei of living ATDC5 cells were labeled with Hoechst 33342 for 30 minutes and incubated for 2 hours in medium with 5 mM Pi. Subsequently, LMS were isolated and stained with a membrane label (FM 4-64) and a label for mineral deposition (calcein). Visualization of DNA and membranes in the crystals showed that the crystals contained whole cells, which seemed to affect the shape of the crystal (Fig. 4A). Furthermore, the nuclei present in the crystals were swollen (arrowhead Fig 4A) or DNA was present in fiber-like structures (arrow Fig. 4A and Fig. 4C-D), which

Chapter 2

38

appeared to be integrated in the HAP crystal. The nuclei of the cells growing on the cell culture plate appeared normal (Fig. 4B). Occasionally calcein stain spots were detected in some cells (data not shown, c.f. Fig. 7)

Figure 3. Large mineralizing structures contain proteins and DNA. Subconfluent ATDC5 cells were incubated for the indicated time points in α-MEM (C, ▲), or in α-MEM + 5 mM Pi (P, ■). Precipitates were isolated from the medium by centrifugating for 20 minutes at 20,000 x g, and solubilized in 60 mM EDTA. Cells remaining on the cell culture dish were scraped off the plate and lysed with an ultrasone disintegrator. A) Proteins from precipitates and total cell lysate (1:100 diluted) were separated by SDS-PAGE and polyacrylamide gels were stained in Coomassie brilliant blue (0.1% w/v) staining solution. B) DNA content was determined with Hoechst 33342. The results are expressed as µg DNA/cell culture flask and represent the mean ± S.E.M. of three independent experiments.


39

Figure 4. Large mineralizing structures contain fragmented cells. Subconfluent ATDC5 cells were labeled with Hoechst 33342 (DNA, blue), and 2 hours incubated in medium with addition of 5 mM Pi. Subsequently, LMS were isolated from the medium. LMS and remaining cells on the cell culture dish were then stained with calcein (mineral deposition, green) and FM4-64 (membrane structure, red) as described in the Materials and Methods section. The structures were visualized by a fluorescence microscope. A) LMS, arrow shows DNA as a fiber-like structure, arrow-head shows swollen nuclei, B) Remaining cells on the cell culture dish. C and D) other experiment wherein LMS were solely stained with Hoechst to clearly demonstrate the fiber-like structure of DNA. A representative image of 3 independent experiments is shown. Effect of dead cells on the formation of large mineralizing structures As LMS contained fragmented cells, we studied whether dead or living cells induced the formation of LMS. Therefore, subconfluent living ATDC5 cells were labeled with Hoechst 33342 for 30 minutes and then 24 hours incubated in serum free α-MEM containing either 10 µM staurosporin or 0.2% sodiumazide to induce cell death. Subsequently, 5 mM Pi was added to the medium and incubated for 2 hours. As shown in Fig. 5A, the amount of precipitated calcium was diminished in staurosporin treated cells, and almost no crystals developed in sodiumazide treated cells. Although staurosporin treated cells were able to nucleate calcium phosphate, they were not able to develop LMS (Fig. 5B). In addition, staurosporin treated cells

Chapter 2

40

did not integrate along the shape of the crystal, but appeared as round condensed dead cells (Fig. 5B), similar to sodiumazide treated cells (Fig. 5C). These data suggest that cells need to be alive in order to form LMS.

Figure 5. Effect of dead cells on the formation of large mineralizing structures. Subconfluent ATDC5 cells were either directly (control), or after 24 hours of pretreatment with either 10 µM staurosporin (staurosporin), or 0.2% sodiumazide (azide), incubated in with addition of 5 mM Pi for 2 hours. Media were centrifuged at 20,000 x g for 20 minutes and precipitates were collected. A) Calcium and DNA were measured as described in the Materials and Methods section. As the treatment to initiate cell death resulted in different amounts of DNA in the medium, the results are expressed as µg calcium/µg DNA and represent the mean ± S.E.M. of three independent experiments. * P< 0.05 with respect to the control value. B and C) fluorescence microscopy of staurosporin- and sodiumazide treated cells incubated for 2 hours with Pi. Subconfluent ATDC5 cells were first labeled with Hoechst 33342 (DNA, blue), then treated with either B) staurosporin or C) sodiumazide, and subsequently incubated in α-MEM + 5 mM Pi for 2 hours. Isolated precipitates were then stained with calcein (crystal structure, green) and FM4-64 (membrane structure, red), and precipitates were visualized by a fluorescence microscope. A representative image of 3 independent experiments is shown.


41

Effect of phosphonoformic acid on the development of large mineralizing structures Subsequently, we investigated whether Pi influx in a living cell is required for the formation of LMS by addition of phosphonoformic acid (PFA). PFA has been shown to inhibit phosphate transport and mineralization in long-term cell culture systems containing serum [33;34]. As shown in Fig. 6A PFA inhibited LMS formation, suggesting that Pi has to enter the cell in order to induce LMS formation. Of note, the amount of precipitated DNA and proteins in the medium was also inhibited by the addition of PFA (data not shown)

Figure 6. The phosphate transporter inhibitor phosphonoformic acid (PFA) inhibits formation of large mineralizing structures. Subconfluent ATDC5 cells were incubated for 2 hours in α-MEM (-), or in α-MEM + 5 mM Pi (Pi), or in α-MEM + 5 mM Pi + 1 mM PFA (PFA + Pi) A) Precipitates were isolated from the medium by centrifugation for 20 minutes at 20,000 x g and calcium was extracted and measured by atomic absorption spectrometry as described in the Materials and Methods section. The results are expressed as µg calcium/cell culture flask and represent the mean ± S.E.M. of three independent experiments. B) Phosphate induces translocation of Annexin V in ATDC5 cells. Cells were scraped off the plate, lysed and separated in a membrane and a cytosolic fraction as described in the materials and methods section. Proteins (20 µg) were separated on SDS-PAGE and the amount of annexin V was determined by western blotting using a rabbit polyclonal annexin V antibody. Results shown are representative for three independent experiments.

Chapter 2

42

Effect of Pi on annexin V translocation on growing cells To investigate whether Pi induces an intracellular effect, the translocation of Annexin V was investigated. Annexin V is a calcium-dependent membrane binding protein [35-37]. In addition, Annexin V has been shown to play a major role in mineralization [25]. Fig. 6B shows that Pi induced a translocation of Annexin V from the cytosol to the membrane in all cells, and this translocation was partly inhibited by PFA. These data indicate that Pi enters and signals the cells to translocate annexin V to the membrane.

Figure 7. Phosphate induces small mineralizing structures inside cells in the presence of serum. ATDC 5 cells were grown for 14 days in DMEM/F12 and subsequently 2 days in α-MEM + 5 mM Pi in the presence of 5% FBS. Non-fixed cells were labeled with Hoechst 33342 (DNA, blue), calcein (crystal structure, green) and Annexin V (PS externalization, red), and visualized by a multiphoton microscope as described in Materials and Methods section. Arrows show apatite crystal within cells, arrow-heads show enlarged apatite crystals in the matrix. Of note, this is a confocal section. A representative image of 3 independent experiments is shown. Small mineralizing structures inside cells in the presence of serum LMS did not develop in the presence of 5% FBS in the medium (data not shown). To see where mineralizing structures initiate in the presence of serum, we


43

visualized the mineralizing structures in the extracellular matrix in a long-term cell culture system (cells grown for 14 days and subsequently 2 days in medium + 5 mM Pi in the presence of 5% FBS). Non-fixed (living) cells were stained with Hoechst 33342 (DNA, blue), calcein (mineral deposition, green) and annexin-V-Alexa 568 (plasma membrane of apoptotic or leaky cells, red). This latter probe was used as a tool to visualize the plasma membrane of apoptotic or leaky cells in the cell culture matrix [38]. As shown in Fig. 7, we observed a number of calcium containing crystals clearly within the cell of a confocal section (arrows Fig. 7, section thickness 1 µm). Rather enlarged apatite crystals were also observed and appeared not surrounded by a membrane (arrow heads Fig. 7). Possibly, these structures were formed in cells. As a result of crystal growth, these cells completely fragmented and the cell membranes disappeared. Calcein labeled spots were not detected in 14 days grown cell cultures without additional Pi. Of note, intracellular visualization of apatite crystals was only possible in leaky cells, since live cells are impermeable for calcein. Taken together, these data suggest that in the presence of serum mineral crystal growth may occur intracellularly. Discussion This study shows that addition of 5 mM phosphate to serum free medium results in a relatively rapid (within 2 hours) formation of large mineralizing structures (LMS) by ATDC5 cells. Generally, to investigate cell-mediated mineralization in vitro, cells are grown until an extracellular matrix is formed. This requires long-term cell culture systems (approximately 10-21 days) in medium containing specific growth factors. During or after the growth period an increase in phosphate (2 - 10 mM) or betaglycerophosphate (2 – 10 mM) for several days is necessary to induce matrix mineralization [21;24;39;40]. However, in these cell cultures mineralizing structures are strongly embedded in the extracellular matrix, which makes it difficult to investigate them by isolation [15;20]. Therefore, we used a culture system without the presence of an extracellular matrix. Furthermore, we shortened the cell culture period by excluding serum, which is known to contain fetuin, from the medium [12;41;42]. Of note, LMS also developed in the medium when cells had developed an extracellular matrix after 14 days (data not shown), which indicates that the rapid formation of LMS is mainly induced by excluding serum from the phosphate-containing medium. The formation of LMS in the absence of serum may provide some fundamental information about the beginning stages of cell-mediated mineralization. The HAP crystals in our serum free short term cell culture model were able to expand to sub millimeter dimensions, which to our knowledge has never been reported before. Our data show that in the LMS, the cells and their nuclei were stretched apart and

Chapter 2

44

integrated in the crystal. For comparison, staurosporin-treated cells may nucleate calcium phosphate in secreted apoptotic bodies [43-45]. Staurosporin-treated cells may also nucleate calcium phosphate by the externalization of phosphatidylserine [5;46;47]. Based on the result of this study the mineralizing structures induced by the staurosporin-treated cells do not resemble LMS and do not contain distorted nuclei (see Fig. 5). This indicates that the formation of LMS is not mediated by late apoptotic cells. If crystal nucleation would start extracellularly, we would not expect the cells to have a stretched morphology. Therefore, the characteristics of the LMS also do not fit in the theory of nucleating non-collagenous phosphoproteins residing in the extracellular matrix [3;48;49]. In our view, the most likely explanation for the formation of the LMS is that various crystal nucleation cores are formed inside the cell and grow into different directions, which will then stretch the cells apart. The presence of multiple mineralization cores inside a cell was observed under circumstances that the mineralization process was attenuated by the inclusion of serum in the medium (Fig. 7). In long-term cell culture systems in the presence of serum, it may be possible that these intracellular crystal containing vesicles are secreted by the cells, as a protective mechanism to remove the crystals [21;50]. Only a small population of cells was shown to be integrated in LMS (approximately 0.5% of the DNA of the total cell population). This suggests that only a subpopulation of cells is competent to develop and integrate into LMS. This finding resembles the observation that in vascular smooth muscle cell culture systems only a subpopulation of cells undergoes osteoblastic differentiation and induces mineralization [9]. In addition, in chondrogenic long term cell culture systems the detection of non homogenous alizarin red positive spots (a dye that detects mineralization by staining it red) shows that mineralization does not occur in all cells [27;51]. This is another indication that a small population of mineralization competent cells induces mineralization. The results of this study also show that the ADTC5 cells need to be alive in order to form LMS, suggesting that the cells need to sequester calcium and phosphate actively. This is in line with the observation that the phosphate transporter inhibitor PFA was able to inhibit the formation of LMS. Also in long term cell culture systems evidence was shown that Pi influx is necessary for the induction of mineralization [52-54]. The possibility that the increased Pi concentration had an effect on Ca levels within the cells was supported by our finding that Pi induced translocation of annexin V from the cytosol to the membrane in the cells on the cell culture dish. Annexin V has been shown to be highly expressed in hypertrophic and mineralizing growth plate cartilage and in bone [55-57]. Furthermore, translocation of annexin V to the membrane has been shown to be calcium dependent and reported to result in the formation of calcium transport channels in a PS rich


45

membrane [25;35;36]. This suggests that Pi may induce calcium transport into the cells, and subsequently crystals may develop intracellularly. In summary, this study shows the cell-mediated rapid formation of LMS in a serum free cell culture system containing additional Pi. Although the conditions in this simplified rapid model system may differ from those in the entire organism, data presented in this study are in our view most compatible with a model in which the mineralization process can be initiated within a cell. This alternative point of view may add to the complexity of the process of mineralization. Furthermore, the benefit of this serum free model is that it may open the possibility to follow the beginning stages of cell-mediated mineralization, which is currently performed in our laboratory. Acknowledgement We thank Hugo S. Wouterse for expert assistance with the atomic absorption spectroscopy. We also thank the Centre for Cell Imaging at the Faculty of Veterinary Medicine, Utrecht University for using the fluorescence- and multiphoton microscope. Reference List 1. Kronenberg HM (2003) Developmental regulation of the growth plate. Nature 423:332-336 2. Linde A, Lundgren T (1995) From serum to the mineral phase. The role of the odontoblast

in calcium transport and mineral formation. Int.J.Dev.Biol. 39:213-222 3. Butler WT, Ritchie H (1995) The nature and functional significance of dentin extracellular

matrix proteins. Int.J.Dev.Biol. 39:169-179 4. Veis A (2005) Materials science. A window on biomineralization. Science 307:1419-1420 5. Speer MY, Giachelli CM (2004) Regulation of cardiovascular calcification.

Cardiovasc.Pathol. 13:63-70 6. Giachelli CM (2005) Inducers and inhibitors of biomineralization: lessons from

pathological calcification. Orthod.Craniofac.Res. 8:229-231 7. Magne D, Julien M, Vinatier C, Merhi-Soussi F, Weiss P, Guicheux J (2005) Cartilage

formation in growth plate and arteries: from physiology to pathology. Bioessays 27:708-716 8. Bobryshev YV (2005) Transdifferentiation of smooth muscle cells into chondrocytes in

atherosclerotic arteries in situ: implications for diffuse intimal calcification. J.Pathol. 205:641-650

9. Demer LL, Tintut Y (2003) Mineral exploration: search for the mechanism of vascular calcification and beyond: the 2003 Jeffrey M. Hoeg Award lecture. Arterioscler.Thromb.Vasc.Biol. 23:1739-1743

10. Triffitt JT, Gebauer U, Ashton BA, Owen ME, Reynolds JJ (1976) Origin of plasma alpha2HS-glycoprotein and its accumulation in bone. Nature 262:226-227

11. Schinke T, Amendt C, Trindl A, Poschke O, Muller-Esterl W, Jahnen-Dechent W (1996) The serum protein alpha2-HS glycoprotein/fetuin inhibits apatite formation in vitro and in

Chapter 2

46

mineralizing calvaria cells. A possible role in mineralization and calcium homeostasis. J.Biol.Chem. 271:20789-20796

12. Schafer C, Heiss A, Schwarz A, Westenfeld R, Ketteler M, Floege J, Muller-Esterl W, Schinke T, Jahnen-Dechent W (2003) The serum protein alpha 2-Heremans-Schmid glycoprotein/fetuin-A is a systemically acting inhibitor of ectopic calcification. J.Clin.Invest 112:357-366

13. Jahnen-Dechent W, Schinke T, Trindl A, Muller-Esterl W, Sablitzky F, Kaiser S, Blessing M (1997) Cloning and targeted deletion of the mouse fetuin gene. J.Biol.Chem. 272:31496-31503


15. Tye CE, Hunter GK, Goldberg HA (2005) Identification of the type I collagen-binding domain of bone sialoprotein and characterization of the mechanism of interaction. J.Biol.Chem. 280:13487-13492



18. Ikeda T, Shirasawa T, Esaki Y, Yoshiki S, Hirokawa K (1993) Osteopontin mRNA is expressed by smooth muscle-derived foam cells in human atherosclerotic lesions of the aorta. J.Clin.Invest 92:2814-2820

19. Wada T, McKee MD, Steitz S, Giachelli CM (1999) Calcification of vascular smooth muscle cell cultures: inhibition by osteopontin. Circ.Res. 84:166-178

20. Anderson HC (1995) Molecular biology of matrix vesicles. Clin.Orthop.Relat Res.266-280 21. Reynolds JL, Joannides AJ, Skepper JN, McNair R, Schurgers LJ, Proudfoot D, Jahnen-

Dechent W, Weissberg PL, Shanahan CM (2004) Human vascular smooth muscle cells undergo vesicle-mediated calcification in response to changes in extracellular calcium and phosphate concentrations: a potential mechanism for accelerated vascular calcification in ESRD. J.Am.Soc.Nephrol. 15:2857-2867





26. Kirsch T, Harrison G, Golub EE, Nah HD (2000) The roles of annexins and types II and X collagen in matrix vesicle-mediated mineralization of growth plate cartilage. J.Biol.Chem. 275:35577-35583



47

28. Kim YJ, Sah RL, Doong JY, Grodzinsky AJ (1988) Fluorometric assay of DNA in cartilage explants using Hoechst 33258. Anal.Biochem. 174:168-176

29. Hale LV, Ma YF, Santerre RF (2000) Semi-quantitative fluorescence analysis of calcein binding as a measurement of in vitro mineralization. Calcif.Tissue Int. 67:80-84

30. Uchimura E, Machida H, Kotobuki N, Kihara T, Kitamura S, Ikeuchi M, Hirose M, Miyake J, Ohgushi H (2003) In-situ visualization and quantification of mineralization of cultured osteogenetic cells. Calcif.Tissue Int. 73:575-583

31. Dunn KW, Young PA (2006) Principles of multiphoton microscopy. Nephron Exp.Nephrol. 103:e33-e40

32. Bennett J, Scott KJ (1971) Quantitative staining of fraction I protein in polyacrylamide gels using Coomassie brillant blue. Anal.Biochem. 43:173-182

33. Montessuit C, Caverzasio J, Bonjour JP (1991) Characterization of a Pi transport system in cartilage matrix vesicles. Potential role in the calcification process. J.Biol.Chem. 266:17791-17797

34. Szczepanska-Konkel M, Yusufi AN, VanScoy M, Webster SK, Dousa TP (1986) Phosphonocarboxylic acids as specific inhibitors of Na+-dependent transport of phosphate across renal brush border membrane. J.Biol.Chem. 261:6375-6383

35. Barwise JL, Walker JH (1996) Annexins II, IV, V and VI relocate in response to rises in intracellular calcium in human foreskin fibroblasts. J.Cell Sci. 109 ( Pt 1):247-255

36. Mohiti J, Caswell AM, Walker JH (1995) Calcium-induced relocation of annexins IV and V in the human osteosarcoma cell line MG-63. Mol.Membr.Biol. 12:321-329

37. Raynal P, Kuijpers G, Rojas E, Pollard HB (1996) A rise in nuclear calcium translocates annexins IV and V to the nuclear envelope. FEBS Lett. 392:263-268

38. Vermes I, Haanen C, Steffens-Nakken H, Reutelingsperger C (1995) A novel assay for apoptosis. Flow cytometric detection of phosphatidylserine expression on early apoptotic cells using fluorescein labelled Annexin V. J.Immunol.Methods 184:39-51

39. Garimella R, Bi X, Camacho N, Sipe JB, Anderson HC (2004) Primary culture of rat growth plate chondrocytes: an in vitro model of growth plate histotype, matrix vesicle biogenesis and mineralization. Bone 34:961-970

40. Magne D, Bluteau G, Faucheux C, Palmer G, Vignes-Colombeix C, Pilet P, Rouillon T, Caverzasio J, Weiss P, Daculsi G, Guicheux J (2003) Phosphate is a specific signal for ATDC5 chondrocyte maturation and apoptosis-associated mineralization: possible implication of apoptosis in the regulation of endochondral ossification. J.Bone Miner.Res. 18:1430-1442






46. Boskey AL, Dick BL (1991) The effect of phosphatidylserine on in vitro hydroxyapatite growth and proliferation. Calcif.Tissue Int. 49:193-196

Chapter 2

48

47. Bratton DL, Fadok VA, Richter DA, Kailey JM, Guthrie LA, Henson PM (1997) Appearance of phosphatidylserine on apoptotic cells requires calcium-mediated nonspecific flip-flop and is enhanced by loss of the aminophospholipid translocase. J.Biol.Chem. 272:26159-26165

48. Alford AI, Hankenson KD (2006) Matricellular proteins: Extracellular modulators of bone development, remodeling, and regeneration. Bone

49. Donley GE, Fitzpatrick LA (1998) Noncollagenous matrix proteins controlling mineralization; possible role in pathologic calcification of vascular tissue. Trends Cardiovasc.Med. 8:199-206

50. Fleckenstein-Grun G, Thimm F, Czirfuzs A, Matyas S, Frey M (1994) Experimental vasoprotection by calcium antagonists against calcium-mediated arteriosclerotic alterations. J.Cardiovasc.Pharmacol. 24 Suppl 2:S75-S84

51. Huitema LF, Vaandrager AB, van Weeren PR, Barneveld A, Helms JB, van de Lest CH (2006) The nitric oxide donor sodium nitroprusside inhibits mineralization in ATDC5 cells. Calcif.Tissue Int. 78:171-177

52. Giachelli CM, Jono S, Shioi A, Nishizawa Y, Mori K, Morii H (2001) Vascular calcification and inorganic phosphate. Am.J.Kidney Dis. 38:S34-S37

53. Wang W, Xu J, Du B, Kirsch T (2005) Role of the progressive ankylosis gene (ank) in cartilage mineralization. Mol.Cell Biol. 25:312-323

54. Yang H, Curinga G, Giachelli CM (2004) Elevated extracellular calcium levels induce smooth muscle cell matrix mineralization in vitro. Kidney Int. 66:2293-2299

55. Crompton MR, Moss SE, Crumpton MJ (1988) Diversity in the lipocortin/calpactin family. Cell 55:1-3

56. Genge BR, Wu LN, Wuthier RE (1989) Identification of phospholipid-dependent calcium-binding proteins as constituents of matrix vesicles. J.Biol.Chem. 264:10917-10921

57. Kirsch T, Swoboda B, Nah H (2000) Activation of annexin II and V expression, terminal differentiation, mineralization and apoptosis in human osteoarthritic cartilage. Osteoarthritis.Cartilage. 8:294-302

Appendix A to Chapter 2

Subcellular localization of mineral crystals in ATDC5 cells.

Leonie F.A. Huitema1,2, Anko M. de Graaff1, P. René van Weeren2, Chris H. A. van de Lest1,2, Ab Barneveld2, J. Bernd Helms1, Arie B. Vaandrager1

1 Department of Biochemistry and Cell Biology, 2 Department of Equine Sciences, Faculty of Veterinary Medicine, and Graduate School Animal Health, Utrecht

University, Utrecht, the Netherlands


50

Introduction As shown in Chapter 2 (Fig. 7), we observed a number of calcium containing crystals within non-fixed ATDC5 cells by confocal microscopy after staining with calcein (mineral deposition) and Annexin-V-Alexa 568 (plasma membrane of apoptotic or leaky cells [1] ). Since calcein is membrane impermeable, we could only detect mineral crystals in apoptotic or leaky cells in that study. Here, we tried to assess whether intracellular mineral crystals are also present in non-apoptotic cells, by looking for calcein staining in non-Annexin V stained cells after permeabilization. Furthermore, we tried to determine in which subcellular compartment the mineral crystals were formed, by investigating a possible co-localization of calcein with antibodies for Golgi (anti-GM130), microtubules (anti-tyrosine tubulin), lysosomes (anti-LAMP-2) and endoplasmatic reticulum (ER) (anti-calnexin) in fixed ATDC5 cells. Materials and methods Materials FM 4-64 was from Invitrogen (Breda, The Netherlands) and Annexin-V-Alexa 568 was from Roche Molecular Biochemicals (Mannheim, Germany). Mouse monoclonal anti-GM130 was from BD Biosciences (Alphen aan den Rijn, The Netherlands), mouse monoclonal anti-tyrosine tubulin was obtained from Sigma (Poole, UK), mouse monoclonal anti-LAMP-2 was from Developmental Studies Hybridoma Bank (Baltimore, MD, USA)) and rabbit polyclonal calnexin was from Stressgen Bioreagents (Michigan, USA). Fluor Save TM reagent was obtained from EMD Biosciences (San Diego, USA). Induction of mineralization in the presence of serum ATDC5 cells were grown for 14 days on coverslips (ø 25 mm) in 6 well plates (ø 35 mm/well) for live cell imaging, or on coverslips (ø 12 mm) in 4 well plates (ø 15.4 mm/well) for imaging of cells after fixation. Subsequently, mineralization was induced by transferring the cells to α-MEM containing 5% FBS, ITS-X, 100 IU/ml penicillin, and 100 µg/ml streptomycin (mineralization medium) with or without 5 mM phosphate at 37°C in a humidified atmosphere of 5% CO2 – 95% air for 24 hours. Detection of mineral crystals in non-fixed cells To visualize calcium containing mineral crystals in leaky, non-fixed, ATDC5 cells, the cells were incubated for at least 1 hour with 5 µg/ml calcein and then 30 minutes with the fluorescent DNA label, Hoechst 33342 (2 µg/ml) in

Subcellular localization of mineral crystals

51

mineralization medium. Subsequently, ATDC5 cells were washed and incubated with FM 4-64 (5 µg/ml) in mineralization medium for at least 2 minutes. To visualize intracellular mineral crystals in living, non-apoptotic, ATDC5 cells, the cells were incubated for 30 minutes with the fluorescent DNA label, Hoechst 33342 (2 µg/ml) and Annexin-V-Alexa 568 (1:50 diluted) in mineralization medium. After incubation, cells were washed 3 times with medium and incubated for at least 20 minutes with Leibovitz’s (L15) medium containing 0.1% saponin and 5 µg/ml calcein and subsequently imaged by multiphoton microscopy as described below. After detection of a mineral crystal in the proximity of a nucleus of a cell not stained with Annexin V, membrane structures of the cells were visualized after 2 min incubation with FM 4-64 (5 µg/ml), which was added to the cells under the microscope. Cells were imaged in a 37°C incubation chamber (solent scientific, Portmouth, UK) mounted on a BioRad Radiance 2100 MP multiphoton microscope (BioRad, Hemel Hempstead, UK) with an 100x oil objective [2]. Annexin-V-Alexa 568 was imaged using a He/Ne laser (λex 543 nm, λem 600/50); FM4-64 was imaged using an Ar laser (λex 488 nm, λem 600/50). Calcein fluorescence was imaged using an Ar laser (λex 488 nm, λem 515/30); Hoechst 33342 fluorescence was imaged using a mode-locked Ti:Sapphire laser (Tsunami, Spectra Physics, Mountain View, CA) pumped by a 10 W solid state laser (Millennia Xs, Spectra Physics, Mountain View, CA) (λex 780 nm, λem 450/80). Subcellular localization of mineral crystals in fixed cells ATDC5 cells were fixed with 95% ethanol for 15 minutes and then permeabilized in 0.5% Triton for 5 minutes. After blocking with phosphate buffered saline (PBS) containing 2% serum albumin (BSA) for 20 minutes, the cells were incubated with the primary antibodies at room temperature for 1 hour (1:200 diluted mouse monoclonal anti-GM130 for Golgi structures, 1:100 diluted mouse monoclonal anti-tyrosine tubulin for microtubules, 1:100 diluted mouse monoclonal anti-lysosomal associated membrane protein, LAMP-2, for lysosomal structures and 1:100 diluted rabbit polyclonal anti-calnexin for ER structures). Subsequently, the cells were washed three times with PBS and incubated with Alexa FluorTM568 anti-mouse IgG or anti-rabbit IgG (1:1000 diluted, Molecular Probes, Leiden, The Netherlands) for 1 hour, and washed three times with PBS. The cells were then incubated 30 minutes in 2 µg/ml Hoechst 33342 and washed 3 times with PBS. After that, the cells were incubated with 5 µg/ml calcein for at least 1 hour and washed 5 times with PBS. After PBS washes, cells were preserved in Fluor SaveTM reagent. Fluorescently labelled cell cultures were visualized by placing the specimen on a BioRad Radiance 2100 MP multiphoton microscope as described above.


52

Figure 1. Subcellular localization of mineral crystals in ATDC5 cells. ATDC5 cells were grown for 14 days in differentiation medium and subsequently 24 hours in mineralization medium with or without addition of 5 mM Pi. Cell cultures were fixed in 95% ethanol and Golgi (A), microtubules (B), lysosomes (C) and ER (D) were visualized with anti-GM130 (Golgi marker), anti-tyrosine-tubulin (microtubules marker), anti-LAMP-2 (lysosomal marker) and anti-calnexin (ER marker) respectively, in combination with Alexa fluor 568 anti-mouse or anti-rabbit IgG (red). Subsequently, the cells were stained with Hoechst 33342 (blue) and calcein (green). Control = mineralization medium without additional Pi; Mineralization = mineralization medium with additional Pi. Arrows show cells filled with calcein labelled mineral crystals


53

Figure 2. Mineral crystals in non-fixed ATDC5 cells. ATDC5 cells were grown for 14 days in differentiation medium and subsequently 24 hours in mineralization medium with addition of 5 mM Pi. A) Living (non-fixed) cells were stained with calcein (green) and Hoechst 33342 (blue) and subsequently with FM4-64 (red). B) To search for intracellular mineral crystal in non-apoptotic ATDC5 cell, cell cultures were first stained with Hoechst 33342 (blue) and Annexin V-Alexa 568 (red), and then cells were permeabilized with 0.1% saponin and stained with 5 µg/ml calcein (green). Notice calcein labeled mineral crystal without co-localization with Annexin V-Alexa 568. C) Same image as B, after subsequent addition of FM 4-64 (red) to stain cell membranes. Arrows shows possible intracellular calcein labeled mineral crystals. This is a representative image of 5 experiments. Results Search for intracellular mineral crystals in a living cell As shown in Fig. 1A, mineral crystals labeled with the membrane-impermeable dye calcein (green) were surrounded by FM 4-64 stained cellular membrane structures, suggesting that they were present inside an apoptotic or leaky cell. Next, we wanted to see whether mineral crystals could also be detected in non-leaky cells by permeabilizing the cells prior to the calcein labeling. To differentiate between living and apoptotic/leaky cells, ATDC5 cells were

first stained with Hoechst 33342 (blue in Fig. 1) and Annexin V-Alexa 568 (red in Fig. 1). Subsequently, live (non-fixed) ATDC5 cells were permeabilized with 0.1% saponin and stained with 5 µg/ml calcein (green in Fig. 1). When a mineral crystal was detected that was possibly inside a cell (i.e. close to a nucleus) and not co-localized with Annexin V (See Fig. 1B, arrows), all cell membranes were subsequently visualized by addition of 5 µg/ml FM 4-64 (red in Fig. 1C). As shown in Fig. 1B and C, it was difficult to clearly determine the outlines of the different ATDC5 cells in the culture, since no clear plasma membranes could be


54

observed, possibly due to the permeabilization procedure. Therefore, we were not able to determine whether mineral crystals were localized inside living cells or present between the cells. Subcellular localization of mineral crystals in fixed cells To analyze in which sub-cellular compartment the calcium containing mineral crystal were formed, organelle-specific antibodies, like anti-GM130 (Golgi)(Fig. 2A), anti-tyrosine tubulin (microtubules) (Fig. 2B), anti-LAMP-2 (lysosomes) (Fig. 2C) and anti-calnexin (ER) (Fig. 2D) were visualized (all markers depicted in red in Fig. 2) and analyzed for co-localization with calcein labeled mineral crystals (green in Fig. 2). As shown in Fig. 2, we could detect clear labelling with calcein and the antibodies of interest in ethanol fixed cells. Of note, in paraformaldehyde fixed cells, we could not detect calcein labelled mineral crystals (data not shown). There was no obvious co-localization between the antibodies and the mineral crystals. However, the localization of some of the mineral crystals, marking the outline of a cell, was suggestive for an intracellular origin of these crystals (arrows Fig. 2). Discussion Although we detected intracellular mineral crystals in leaky, presumably apoptotic, ATDC5 cells labeled with FM 4-64, as also observed previously with annexin V (chapter 2 Fig. 7), we could not clearly demonstrate an intracellular mineral crystal in non-apoptotic ATDC5 cells after permeabilization with saponin. Possibly, saponin did not properly permeabilized the internal membranes in the ATDC5 cells [3]. However, it was observed that after addition of saponin, the membrane impermeable DNA stain propidium iodide stained all the nuclei in the cell culture (data not shown). Alternatively, it may be possible that once a crystal is formed, the cell quickly dies and becomes leaky, since mineral crystals are possibly very toxic to the cell [4], or that intracellularly formed mineral crystals are immediately secreted by the cells, as a protective mechanism to remove the crystals. If these actions occur, they may occur quickly after the initial crystal formation it would explain that it is difficult to detect a mineral crystal inside in a living cell. We did not observe a co-localization of calcein with GM130 (Golgi), tyrosine tubulin (microtubules), LAMP-2 (lysosomes) and calnexin (ER). Possibly, the mineral crystals were already expanding over the specified compartment, precluding visualization of the compartment by the antibodies. Alternatively, the crystal growth is initiated in another compartment, not probed by our antibodies, for example in mitochondria [4].


55

Reference List 1. Vermes I, Haanen C, Steffens-Nakken H, Reutelingsperger C (1995) A novel assay for

apoptosis. Flow cytometric detection of phosphatidylserine expression on early apoptotic cells using fluorescein labelled Annexin V. J.Immunol.Methods 184:39-51

2. Dunn KW, Young PA (2006) Principles of multiphoton microscopy. Nephron Exp.Nephrol. 103:e33-e40

3. Wassler M, Jonasson I, Persson R, Fries E (1987) Differential permeabilization of membranes by saponin treatment of isolated rat hepatocytes. Release of secretory proteins. Biochem.J 247:407-415

4. Shapiro IM, Adams CS, Freeman T, Srinivas V (2005) Fate of the hypertrophic chondrocyte: microenvironmental perspectives on apoptosis and survival in the epiphyseal growth plate. Birth Defects Res.C.Embryo.Today 75:330-339

Appendix B to Chapter 2

Inorganic phosphate does not induce a major effect on the cytosolic free calcium concentration in ATDC5

cells Leonie F.A. Huitema1,2, Remco H.S. Westerink3, P. René van Weeren2, Chris H. A. van de Lest1,2, Ab Barneveld2, J. Bernd Helms1, Arie B. Vaandrager1

1 Department of Biochemistry and Cell Biology, 2 Department of Equine Sciences, 3 Institute for Risk Assessment Sciences, Faculty of Veterinary Medicine, and Graduate School Animal Health, Utrecht University, Utrecht, the Netherlands


58

Introduction As described in chapter 2 (Fig. 6B), we observed that addition of 5 mM inorganic phosphate (Pi) induces a translocation of Annexin V from the cytosol to the membrane fraction in ATDC5 cells. Annexin V is a calcium-dependent membrane binding protein, that has been shown to relocate to the cell membrane in response to an elevation of the intracellular concentration of calcium [1-3]. This suggests that Pi may induce the observed translocation of Annexin V by eliciting an increase in the cytosolic calcium (Ca2+) concentration. To test this possibility, we determined the effect of Pi on the cytosolic free calcium concentration in ATDC5 cells using the high affinity dye Fura-2-AM. Materials and methods Materials Leibovitz’s (L-15) medium was obtained from Gibco BRL (Grand Island, NY). Fura-2 acetoxylmethyl (AM) was from Invitrogen, Molecular Probes (Breda, The Netherlands). All other chemicals were obtained from Sigma (Poole, UK). Calcium- imaging ATCD5 cells, grown in DMEM/F12 medium, with an initial density of 15,000 cells/cm2 on glass coverslips (ø 12 mm) for 1 day, were incubated with 5 µM of the calcium indicator Fura-2-AM in L-15 medium at room temperature for 30 min. After incubation, cells were washed with L15 medium and kept in 1 ml L15 medium at room temperature for 15 min to allow intracellular deesterification of Fura-2-AM. After deesterification, cells were placed on the stage of an Axiovert 35M microscope (Zeiss) equipped with a monochromator (TILL Photonics Polychrome IV lamp) that permits switching between 340 nm and 380 nm wavelengths used for excitation. The fluorescence was collected at a wavelength of 510 nm with a TILL Photonics Image digital camera (TILL photonics GmBH, Gräfelfing, Germany). Digital camera and monochromator were controlled by imaging software (TILL VISion, v4.01), which was also used for data collection and processing. F340/F380 ratio was used for qualitative description of changes in the cytosolic calcium concentration. Exposure protocol After a transmission image was made, collection of fluorescence ratio-images started. The first 5 minutes of the recording (4 ratio-frames/min) served as control; cells with abnormal levels of intracellular Ca2+ were discarded from further analysis. Subsequently, 1 ml of L15 medium, either with or without 10 mM phosphate was added to the cells (final additional Pi concentration was 0 or 5 mM),

Cytosolic free calcium in ATDC5 cells

59

and recordings continued at room temperature (~23°C) for 120 min (3 ratio-frames/min). At the end of the recording a final transmission image was made. Experiments with and without Pi were performed twice (a total amount of 32 individual cells with 5 mM Pi and a total amount 37 individual cells without Pi). Then, ionomycin (5 µM) and EDTA (17 mM) were added to obtain the maximum and minimum F340/F380 ratio, respectively. Results ATDC5 cells grown for one day were labelled with Fura-2 and subsequently treated with serum free L15 medium either with or without additional 5 mM Pi for 2 hours. As shown in Fig. 1A, a small initial rise in Ca2+ concentration was observed in cells with additional Pi and control cells (without additional Pi). This may represent an artefact due to addition of fresh medium. The addition of Pi to the cells did not lead to an increase of cytosolic Ca2+. There was also not much variation between the individual cells in experiments with Pi (0.58 ± 0.05 F340/F380) or without Pi (0.6 ± 0.07 F340/F380), though the average F340/F380 base line (Fig. 1A, 0-5 minutes) of cells to be treated with L15 medium without additional Pi already started at a higher value compared to cells to be treated with Pi. A characteristic of ATDC5 cells is the occurrence of oscillations in the intracellular Ca2+ concentration in 15-30% of the cells during the recording period of 2 hours (see Fig 1B for example). However, as with the basal cytosolic Ca2+ concentration, differences between cells with additional Pi and without additional Pi could not be detected for the occurrence, frequency, duration and amplitude of the oscillations. Discussion This study shows that addition of 5 mM Pi does not induce an increase in the cytosolic free calcium concentration in ATDC5 cells. For practical reasons, the cells were incubated in a bicarbonate-free L15 medium, while in previous experiments (chapter 2), the cells were incubated in a bicarbonate buffered medium (α-MEM). Therefore, we cannot formally exclude the possibility that the Pi may have had an effect on Ca2+ levels in the previous studies using α-MEM. However, this is not very likely, since we observed that the large mineralizing structures (LMS) were formed in L15 medium in a similar way as in α-MEM (L.F.A. Huitema, unpublished observations). It may be possible that the Pi-induced Annexin V translocation from the cytoplasm to the membrane in ATDC5 cells reported in chapter 2 is not mediated by an increase in the cytosolic Ca2+ concentration. However, to our knowledge, no Ca2+ independent annexin V membrane binding has been described. Possibly, Pi activates another signal cascade leading to Annexin V translocation not involving


60

Ca2+ influx. Alternatively, a small Pi-induced Ca2+ influx, either from intracellular stores or from the extracellular environment, may cause a small local rise in Ca2+ close to the membrane before the calcium is recruited to specific intracellular stores, like for example the ER or mitochondria [4;5]. This small and local increase in Ca2+ may lead to a translocation of Annexin V without causing a general increase in cytosolic calcium. Therefore, it would be worthwhile to investigate the effect of Pi on Ca2+ handling by the ER or the mitochondria. In chapter 2 we reported that 0.5% cells of the total cell population are integrated in large mineralizing structures. During these experiments we possibly missed these cells in the image field as we only imaged approximately 16 cells per experiment.

Cytosolic free calcium in ATDC5 cells

61

Figure 1. Pi does not induce an increase in the cytosolic Ca2+ concentration in ATDC5 cells. ATCD5 cells, grown on glass coverslips for 1 day, were labelled with 5 µM Fura-2-AM. Cells were exposed to L-15 medium with or without additional 5 mM Pi 5 minutes after the start of the recording, as indicated with an arrow. Data were obtained by averaging fluorescence within small regions-of-interest within the cytoplasm of several cells in the image field. The Fura-2 data are expressed as the ratio of fluorescence excited at 340 and 380 nm. A) Average of F340/F380 data (F340/F380 average) from 32 cells treated with 5 mM Pi (black thick line) and 37 cells treated without additional Pi (black thin line). B) Example of individual cells treated without addition Pi with one (black line) or several (grey line) oscillations.


62

Reference List 1. Barwise JL, Walker JH (1996) Annexins II, IV, V and VI relocate in response to rises in

intracellular calcium in human foreskin fibroblasts. J.Cell Sci. 109 ( Pt 1):247-255 2. Mohiti J, Caswell AM, Walker JH (1995) Calcium-induced relocation of annexins IV and V

in the human osteosarcoma cell line MG-63. Mol.Membr.Biol. 12:321-329 3. Raynal P, Kuijpers G, Rojas E, Pollard HB (1996) A rise in nuclear calcium translocates

annexins IV and V to the nuclear envelope. FEBS Lett. 392:263-268 4. Shapiro IM, Adams CS, Freeman T, Srinivas V (2005) Fate of the hypertrophic

chondrocyte: microenvironmental perspectives on apoptosis and survival in the epiphyseal growth plate. Birth Defects Res.C.Embryo.Today 75:330-339

5. Hajnoczky G, Davies E, Madesh M (2003) Calcium signaling and apoptosis. Biochem.Biophys.Res Commun. 304:445-454

Chapter 3

Soluble factors released by ATDC5 cells affect the formation of calcium phosphate crystals.

Leonie F.A. Huitema1,2, P. René van Weeren2, Bas W.M. van Balkom3,4, Chris H.A. van de Lest1,2, Ab Barneveld2, J. Bernd Helms1, Arie B. Vaandrager1

1 Department of Biochemistry and Cell Biology, 2 Department of Equine Sciences, Faculty of Veterinary Medicine, and Graduate School Animal Health, 3Institute of Veterinary Research, 4Department of Biomolecular Mass Spectrometry, Utrecht


Submitted

Chapter 3

64

Abstract During biomineralization the organism controls the nature, orientation, size and shape of the mineral phase. The aim of this study was to investigate whether proteins or vesicles that are constitutively released by growing ATDC5 cells have the ability to affect the formation of the calcium phosphate crystal. Therefore, subconfluent cultured ATDC5 cells were incubated for 1 hour in medium without serum. Subsequently, medium was harvested and incubated for 24 hours in the presence of additional Pi. This resulted in the formation of flat mineralizing structures (FMS), consisting of complex irregularly shaped flat crystals, which occasionally contained fiber-like structures (~ 40 µm in size). Without pre-incubation of medium with cells, only small punctate (dot like) calcium phosphate precipitates were observed. The formation of FMS was shown to be caused by soluble factors released by subconfluent ATDC5 cells. Proteomic analysis by mass spectrometry showed that FMS contained a specific set intracellular proteins, serum proteins, and extracellular matrix proteins. Bulk cytosolic proteins derived from homogenized cells or serum proteins did, however, not induce the formation of FMS. Therefore, the formation of FMS seems to be caused by specific soluble factors constitutively released by ATDC5 cells. This in vitro model system can be used as a tool to identify factors in affecting the structure of the biomineral phase.

Soluble factors affect the formation of calcium phosphate crystals

65

Introduction Natural materials such as bones and teeth consist of biocomposites with well-organized and assembled hydroxyapatite (HAP) biomaterials [1-3]. Formation of these biominerals is an essential requirement for the development of hard tissues, which provide mechanical support to vertebrates. During biomineralization, the organism controls the nature, orientation, size and shape of the mineral phase by creating closed compartments or defined channels in which crystals form [3]. Although the exact mechanism of crystal nucleation or inhibition remains unknown, it has been postulated that an interaction occurs between mineral nuclei and soluble proteins, which subsequently influences crystal growth [4]. Generally, biological fluid is metastable with respect to calcium phosphate concentration; i.e., it is below a level of saturation needed for spontaneous precipitation, but above the level of saturation to support crystal growth after the initial nucleus has formed [5;6]. Therefore, it is necessary that the organism exert a great degree of control over nucleation and growth of mineral particles. Numerous proteins isolated from bone and teeth have been proposed to function as either nucleators or inhibitors of HAP formation of which non-collagenous phosphoproteins have been investigated extensively [7-10]. A number of non-collagenous phosphoproteins have been shown to act either as nucleators or inhibitors of HAP formation in dentinogenis or osteogenesis [1]. In vitro studies have shown that some phosphoproteins, like bone sialoprotein (BSP), can nucleate the formation of apatite crystals [8;11;12]. Other phosphoproteins, like osteopontin, inhibit mineralization [7;13]. Matrix vesicles (MVs) have also been shown to act as nucleators of HAP formation [14]. MVs are cell-derived extracellular membrane enclosed particles, about 0.1- 1 µm in size, that initiate mineralization by the action of MV-associated phosphatases, calcium-binding phospholipids and proteins [14]. MVs have been isolated from cultures of mineralizing osteoblasts, chondrocytes and vascular smooth muscle cells [15-17]. The aim of this study was to investigate whether proteins or vesicles that are constitutively released by growing ATDC5 cells have the ability to nucleate and/or to remodel calcium phosphate. The results obtained from this study show that ATDC5 cells release soluble factors, which affect the formation of calcium phosphate crystals. Materials and methods Materials Cell culture plastics were purchased from Greiner bio-one (Frickenhausen, Germany) and Nunc Inc (Rochester, NY). Fetal bovine serum (FBS), antibiotics, and Insulin-Transferrin-Selenium-X (ITS-X) were obtained from Gibco BRL

Chapter 3

66

(Grand Island, NY). α-minimum essential medium (α-MEM) was from Biochrom AG (Berlin, Germany). All other chemicals were obtained from Sigma (Poole, UK). Cell culture ATDC5 cells (Riken Cell Bank, Tsukuba, Japan) were cultured and maintained in a 1:1 mixture of Dulbecco’s modified Eagle medium and Ham’s F12 medium (DMEM/F12) containing 5% FBS, 10 µg/ml human transferrin, 30 nM sodium selenite, 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2 – 95% air. For induction of differentiation, ATDC5 cells were cultured in 175 cm2 flasks in the same medium additionally containing 10 µg/ml insulin from ITS-X (Sigma). Fibronectin heterozygous and -knockout fibroblast cells derived from mouse embryonic stem cells were a generous gift from Dr. Mosher [18]. These cells were cultured in DMEM containing 10% FBS, 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2 – 95% air. Cells were subcultured once a week using trypsin/EDTA solution and with an initial density of 15,000 cells/cm2. Medium was replaced every 2 days. Precipitation of calcium phosphate crystals At day 4, subconfluent ATDC5 cells were washed 3 times with phosphate buffered saline (PBS) and incubated for one hour in α-MEM (25 ml/ 175 cm2 flask), containing 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37°C 5 % CO2. Subsequently, this ‘conditioned’ medium was harvested, transferred to a new culture flask and incubated for 24 hours at 37°C 5% CO2 in the presence of 5 mM Pi (obtained from dilution of a 1 M NaH2PO4/Na2HPO4 pH 7.4 stock solution). Alternatively, conditioned medium was first separated in a fraction that contained insoluble proteins and a fraction containing soluble proteins by centrifugation at 100,000 x g in an OptimaTM LE-80 K Ultracentrifuge (Beckman CoulterTM, U.S.A.) for 1 hour. After the incubation period in the presence of additional Pi, precipitates were collected by centrifugation at 20,000 x g in a Sorvall centrifuge RC 5B plus for 20 minutes. Pellets were washed with MilliQ water and collected by centrifugation in a cooled (4°C) eppendorf centrifuge maximum speed (15,800 x g). Infrared spectroscopy Precipitates were washed 3 times in MilliQ water and dehydrated in a vacuum centrifuge (speedvac concentrator SVC 100H, Savant) overnight. Subsequently, precipitates were prepared as KBr pellets for infrared (IR) spectroscopic analysis.


67

Spectra were recorded on a Perkin-Elmer 2000 Fourier transform infrared (FT-IR) spectrometer by accumulating 16 scans over the wavenumber range of 4000-400 cm-1 with 4 cm-1 optical resolution. For comparison, HAP (Bio-Rad laboratories, Richmond) was also analyzed. Visualization of calcium phosphate crystals Precipitates were suspended in 25 µl MilliQ water and visualized with a Leica DMRA fluorescence microscope with a polarization filter and an additional photomatrix coolsnap CCD camera (Leica Microsystems BV, Rijswijk, The Netherlands). Protein analysis Soluble proteins in the conditioned medium were precipitated with 10% TCA. Briefly, a final concentration of 10% TCA was added to the medium and incubated on ice for 1 hour. Subsequently, proteins were collected by centrifugation in a cooled (4°C) eppendorf centrifuge at maximum speed (15800 x g) for 15 minutes. Pellets were washed twice with ice cold acetone. Proteins in the calcium phosphate precipitates and TCA precipitated proteins from the medium were dissolved in 20 µl sodium dodecyl sulfate polyacrylamide gel electrophoreses (SDS-PAGE) sample buffer (final concentration; 50 mM Tris, 2% SDS, 10% glycerol, 2.5% β-mercaptoethanol and 0.02% bromophenol blue, pH 6.8) containing 60 mM EDTA, separated by SDS-PAGE and silver stained. To this end, the gel was fixed for 1 hour in 40% (v/v) ethanol with 10% (v/v) acetic acid and washed (20 minutes) twice in 30% (v/v) ethanol and once (20 minutes) in MilliQ water. Subsequently, the gel was sensitized for 1 minute in 0.02% (w/v) sodium thiosulphate and washed (20 seconds) three times with MilliQ water. The gel was then incubated for 20 minutes in 0.1% (w/v) silver nitrate solution at 4°C and washed (20 seconds) 3 times with MilliQ water. The gel was developed in 1% (w/v) sodium carbonate containing 0.017% (v/v) formaldehyde solution. When protein bands were clearly visible, the staining was terminated with 5% (v/v) acetic acid. Protein identification by mass spectrometry Proteins from calcium phosphate precipitates were dissolved in SDS-PAGE sample buffer containing 60 mM EDTA, separated by SDS-PAGE and stained with Coomassie brilliant blue R250 (0.1% w/v) staining solution [19]. Bands that were consistently present were considered to be of interest and were excised. Excised bands were washed with water and subjected to in-gel tryptic digestion as described previously [20] with one minor modification, namely that gel pieces were submerged in 50 µl of ammoniumbicarbonate buffer. After o/n digestion at

Chapter 3

68

37ºC, 10 µl of the supernatant was analyzed by nanoflow-LC tandem mass spectrometry using an Agilent 1100 HPLC (Agilent Technologies) to an LTQ ion trap mass spectrometer (Thermo Electron, Bremen, Germany), which was operated as described previously [21]. For all protein band analysis *.dta files were created using Bioworks 3.1 software (Thermo Electron, Bremen, Germany) which were converted to a single peak list in Mascot generic format using in-house developed software. For protein identification, peak lists were submitted to the Mascot version 2.0 search software [22] (Matrix Science, London, UK). The allowed peptide mass tolerance was 0.6 Da, fragment mass tolerance of 0.6 Da, and allowing for 2 missed cleavages, to search mouse IPI (version 3.10) [23] or MSDB (version 20050701) databases [24]. Modifications included in the search were carbamidomethyl at cysteine residues (as a fixed modification) and oxidation at methionine residues (variable). Results ATDC5 cells affect the structure of calcium phosphate crystals To investigate whether proteins or vesicles released constitutively by ATDC5 cells have the ability to affect the formation of calcium phosphate crystals, cells were incubated for 1 hour in serum free medium. Subsequently, this conditioned medium was harvested, transferred to a new cell culture flask and incubated for 24 hours with additional Pi. Fig. 1A shows that this resulted in the formation of flat mineralizing structures (FMS), consisting of complex irregularly shaped flat crystals, which occasionally contained fiber-like structures (~ 40 µm in size). In the negative control (non-conditioned medium containing additional Pi) relatively small punctate (dot like) calcium phosphate precipitates were observed. Occasionally these structures developed into punctate aggregated small structures (~ 3 µm in size) (Fig. 1B). These data suggest that factors released from ATDC5 cells affect the structure of calcium phosphate crystals. Infra red spectra of flat mineralizing structures To determine whether the factors released by ATDC5 cells change the mineral phase of the calcium phosphate crystal, FMS were analyzed by infrared spectroscopy. Fig. 2A shows the spectrum of HAP with two characteristic phosphate absorption doublets at 1091-1029 cm-1 and 602-563 cm-1. These doublets were also detected in the spectra of small calcium phosphate crystals from the negative control (non-conditioned medium containing additional Pi) (Fig. 2C) and in the spectra of FMS (Fig. 2B). These data indicate that the factors released by ATDC5 cells do not change the mineral phase of the calcium phosphate crystal.


69

Figure 1. Formation of mineralizing structures by soluble factors released by ATDC5 cells. Subconfluent ATDC5 cells were incubated in α-MEM for 1 hour. Subsequently, the conditioned medium was harvested, and transferred to a new culture flask. 5 mM Pi was then added and the medium was incubated for 24 hours. After that, medium was centrifuged for 20 minutes at 20,000 x g and the pellet was washed and analysed by polarized light microscopy. A) Precipitates from the conditioned medium. Notice the complex irregular shaped flat crystals, which occasionally contain fiber-like structures. B) Precipitates from the non-conditioned medium. C and D) Conditioned medium was separated in a fraction that contained insoluble proteins and a fraction containing soluble proteins. Subsequently, both fractions were incubated in medium containing additional Pi for 24 hours. C) Precipitates formed in the soluble protein fraction, D) Precipitates formed in the insoluble protein fraction. A representative image of 3 independent experiments is shown.

Chapter 3

70

Figure 2. Infrared spectroscopic analysis of mineralizing structures. A) Hydroxyapatite; B and C) subconfluent ATDC5 cells were incubated in α-MEM for 1 hour. Subsequently, the conditioned medium was harvested, and transferred to a new culture flask. 5 mM Pi was then added and the medium was incubated for 24 hours. After that, medium was centrifuged for 20 minutes at 20,000 x g and the pellet was washed and analysed by infrared spectroscopy. Precipitates were analyzed as KBr discs. B) FMS obtained from conditioned medium, C) calcium phosphate crystals obtained from the negative control (non-conditioned medium). The absorption doublets at the 1091-1029 cm-1 and 602-563 cm-1 are characteristic of HAP mineral. Results shown are representative for three independent experiments.


71

Soluble factors cause the formation of flat mineralizing structures To assess whether the formation of FMS was caused by soluble factors, conditioned medium was separated in a fraction containing insoluble factors and a fraction containing soluble factors. Subsequently, both fractions were incubated for 24 hours in medium containing additional Pi. Fig. 1 C-D shows that FMS only developed in the fraction with soluble factors. As shown in Fig. 3 the amount of proteins in the insoluble fraction (Fig. 3, Lane B) was much smaller compared to the amount of proteins in the soluble fraction (Fig. 3, Lane A). Analysis of proteins in flat mineralizing structures To investigate which specific soluble proteins were enriched in FMS, isolated FMS were dissolved in 60 mM EDTA, separated by SDS-PAGE and silver stained (Fig. 3, lane C1). A subset of proteins in the supernatant from conditioned medium was incorporated in FMS (Fig. 3, compare Lanes A, C1 and D). A similar protein pattern was observed in the lanes showing proteins that were not incorporated in FMS (Fig. 3, Lane D) and proteins in FMS (Fig. 3, Lane C1). To identify the proteins that may affect the structure of the calcium phosphate crystal, 9 bands from Lane C2 (Fig. 3) were cut out from the Coomassie stained gel and digested with trypsin. Resulting peptides were subjected to mass spectrometry. As shown in Table 1, FMS contain intracellular proteins, e.g. pyruvate kinase 3 and elongation factor 1, as well as extracellular matrix proteins, e.g. fibronectin, and serum proteins, e.g. fetuin and serum albumin (Fig. 3, Lane E). Bovine fetuin has a theoretical molecular mass of 39 kD (Table 1). However, on the polyacrylamide gel fetuin behaved as a 80 kD protein, suggesting that fetuin forms a stable dimer [25]. Effect of serum and cellular proteins on the structure of calcium phosphate crystals Since serum proteins are present in FMS, there is a possibility that the effect on the calcium phosphate crystal structure is caused by residual serum proteins still present in the conditioned medium, although the cells had been washed 3 times with PBS. Therefore, the total amount of proteins in the conditioned medium and serum was determined. The total amount of proteins in the conditioned medium was measured to be 5-10 µg per 25 ml. Bovine serum (containing 10 µg protein) incubated in 25 ml medium with additional 5 mM Pi for 24 hours resulted in the formation of relatively small punctate (dot like) structures, similar to the structure of calcium phosphate precipitates in non-conditioned medium (data not shown, e.g. Fig. 1B). Next, it was investigated whether a cytosolic fraction derived from a total cell homogenate (containing 10 µg proteins) affected the structure of the calcium phosphate crystals. Like serum, diluted cell homogenate containing soluble

Chapter 3

72

proteins did not affect the structure of the calcium phosphate crystal (data not shown, e.g. Fig. 1B).

Figure 3. SDS-polyacrylamide gel electrophoresis analyses of proteins in flat mineralizing structures. Lane A to D), subconfluent ATDC5 cells were incubated in α-MEM for 1 hour. Subsequently, medium was harvested and separated in a fraction that contained soluble proteins (Lane A) and a fraction containing insoluble proteins (Lane B). Then, 5 mM Pi was added to the medium from the soluble fraction and incubated for 24 hours. Subsequently, medium was centrifuged for 20 minutes at 20,000 x g. The pellet (Lane C1) contained the FMS. The supernatant, (Lane D) contained proteins that did not integrate in FMS. Lane E) 10 µg serum proteins. Proteins were precipitated using 10% TCA (Lane A, B and D). Polyacrylamide gels were silver stained. A representative gel of three independent experiments is shown. Proteins in FMS were analysed by proteomics (Lane C2). Therefore, the polyacrylamide gel was stained with coomassie Brilliant blue and 9 bands were cut out from the gel, digested with trypsin and analyzed by mass spectrometry.

Table 1. Mass spectrometric analyses of proteins present in FMS (compare Fig. 3, Lane C2) a

Band Acc. Number* Protein name MW Mascot Score Peptides Species Localization

1 IPI00113539 Fibronectin Precursor 275,968 201 8 Mouse Extracellular

2 IPI00466069 Elongation factor 2 96,091

166 4 Mouse Intracellular

3 IPI00229080 Heat shock protein 84b 83,571 145 5 Mouse Intracellular

4 A35714 Fetuin precursor 39,139 120 5 Bovine Serum

5 AAN17824 Bovine Albumin 71274 753 20 Bovine Serum

6 IPI00407130 Pyruvate kinase 3 58,536 143 5 Mouse Intracellular

7 IPI00110753 Tubulin alpha-1 chain 50,788 206 5 Mouse Intracellular

8 IPI00307837 Elongation factor 1-alpha 1 50,623 144 4 Mouse Intracellular

9 IPI00110850 Actin, cytoplasmic 1 42,052 214 4 Mouse Intracellular

* SwissProt or mouse IPI accession number a Bands that were consistently present in FMS in the polyacrylamide gel were considered to be of interest. These bands were excised and subsequently analysed by mass spectrometry.

Effect of exclusion of fibronectin on the structure of calcium phosphate crystals Fibronectin was shown to be present in FMS (Fig. 3, Lane C2 and Table 1). Therefore, the role of fibronectin on FMS formation was investigated. Fibronectin heterozygous and knockout fibroblast cells were incubated for 1 hour in serum free medium. Subsequently, medium was harvested, transferred to a new cell culture flask, and incubated for 24 hours containing additional Pi (i.e. the same experimental conditions as used in the case of the ATDC5 cells). As shown in Fig. 4A and B complex flat shaped crystals, occasionally containing fiber-like structures, were observed in fibronectin heterozygous and knockout cells, similar to the FMS formed from conditioned medium by ATDC5 cells (Fig. 1A).

Figure 4. Effect of fibronectin fibroblast knockout cells on the structure of the calcium phosphate crystal. Subconfluent fibroblast cells were incubated in α-MEM for 1 hour. Subsequently, the conditioned medium was harvested, and transferred to a new culture flask. 5 mM Pi was then added and the medium was incubated for 24 hours. After that, medium was centrifuged for 20 minutes at 20,000 x g and the pellet was washed and analysed by polarized light microscopy. A) Fibronectin heterozygous cells, and B) fibronectin knockout cells. A representative image of 3 independent experiments is shown. Discussion This study shows that subconfluent ATDC5 cells constitutively release soluble factors, which can affect the structure of calcium phosphate crystals, resulting in the formation of FMS. Serum is known to contain mineralization inhibitors such as fetuin, which has a high affinity for the mineral phase and could therefore interfere with the results (e.g. formation of FMS). Although the ATDC5 cells were 3 times washed with 25 ml PBS, there were still some residual serum proteins present in the conditioned medium and these proteins (fetuin and serum albumin) appeared to be the most prominent proteins in FMS. It is possible that serum proteins non-specifically stick to the cells, making it difficult to wash them away. Alternatively, it has been shown that e.g. vascular smooth muscle cells take up serum proteins, like fetuin. Subsequently, fetuin was secreted when the cells are surrounded by a


75

medium deficient in fetuin [26]. It may therefore be possible that the ATDC5 cells secrete serum proteins during the 1-hour incubation period in the medium. However, serum proteins by themselves were neither able to form FMS nor able to affect the structure of calcium phosphate crystals. This indicates that the formation of FMS is performed by cell specific factors. Based on the silver stained gel (Fig. 3), a similar protein pattern was observed in the supernatant fraction of the conditioned medium and in the FMS fraction. This may render the protein specificity of the interaction with FMS questionable. It is well known that crystalline HAP is able to aspecifically bind proteins as it is used for purifying proteins [27]. However, cytosolic proteins derived from homogenized cells at the same concentrations as present in the conditioned medium did not affect the structure of the calcium phosphate precipitate. This suggests that specific factors secreted by ATDC5 cells can affect the structure of the calcium phosphate crystal rather than cytosolic proteins released from leaky cells. Fibronectin was identified in FMS. Fibronectin is a secreted non-collagenous protein, known to be a calcium phosphate remodeling protein [28;29]. However, we observed that fibronectin (-/-) fibroblastic cells were also able to form FMS. These data suggest that fibronectin is not a limiting factor for the formation of FMS. Furthermore, the fact that fibroblast cells can also form FMS, suggests that the formation of FMS is not specific for ATDC5 cells. Elongation factor I alpha and HSP 84 were also identified in FMS. These cytosolic proteins have been reported to interact with aggregates and to reduce the rate of aggregate formation in PolyQ diseases [30]. The aggregates in these diseases form as a result of translation of mutant genes containing an abnormal length of PolyQ repeats, such as Huntington’s disease [31]. In addition, actin and tubulin were found in the aggregates, but these have been reported to be rather aspecific [32]. However, these proteins (elongation factor I alpha, HSP 84, actin and tubulin) are not likely to play a role in the formation of FMS, since cytosolic proteins derived from homogenized cells at a concentration as found in the conditioned medium were shown not to form FMS. The formation of FMS was observed in conditioned medium without cells. We reported previously that subconfluent ATDC5 cells incubated for 2 hours in medium with additional 5 mM Pi without the presence of serum resulted in the formation of large mineralizing structures (LMS) (chapter 2). In this previous study, calcium phosphate precipitation did not occur in the negative control medium. Furthermore, LMS were shown to contain whole cells and were larger (sub millimetre size range) than FMS (approximately 40 µm in size). Possibly, factors involved in FMS formation play also a structural role in LMS formation. Taken together, data of this study show that soluble factors released by ATDC5 cells can affect the structure of the calcium phosphate crystal. We were not able to

Chapter 3

76

identify a specific factor responsible for affecting the structure of the HAP crystal, but an organism might control the nature, orientation and size of a HAP crystal by a combination of specific proteins available. This may then result in the formation of a specific hard biomineral tissue. This in vitro model system may be useful as a tool to identify which specific proteins can affect the structure of the calcium phosphate crystal. Acknowledgement We thank Tom Visser from the department of inorganic chemistry and catalyses at the Debye Institute, Utrecht University for expert assistance on the Fourier transform infrared spectrometer. We also thank the Centre for Cell Imaging at the Faculty of Veterinary Medicine, Utrecht University for using the fluorescence microscope and Dr. Mosher from the department of Pathology and Laboratory Medicine and Medicine, University of Wisconsin, Madison for the fibronectin knockout cells. Reference List 1. Butler WT, Ritchie H (1995) The nature and functional significance of dentin extracellular

matrix proteins. Int.J.Dev.Biol. 39:169-179 2. Kronenberg HM (2003) Developmental regulation of the growth plate. Nature 423:332-336 3. Veis A (2005) Materials science. A window on biomineralization. Science 307:1419-1420 4. He G, Gajjeraman S, Schultz D, Cookson D, Qin C, Butler WT, Hao J, George A (2005)

Spatially and temporally controlled biomineralization is facilitated by interaction between self-assembled dentin matrix protein 1 and calcium phosphate nuclei in solution. Biochemistry 44:16140-16148

5. Blumenthal NC, Posner AS (1973) Hydroxyapatite: mechanism of formation and properties. Calcif.Tissue Res. 13:235-243

6. Eanes ED, Gillessen IH, Posner AS (1965) Intermediate states in the precipitation of hydroxyapatite. Nature 208:365-367

7. Giachelli CM (2005) Inducers and inhibitors of biomineralization: lessons from pathological calcification. Orthod.Craniofac.Res. 8:229-231


9. Hunter GK, Goldberg HA (1993) Nucleation of hydroxyapatite by bone sialoprotein. Proc.Natl.Acad.Sci.U.S.A 90:8562-8565

10. Hunter GK, Goldberg HA (1994) Modulation of crystal formation by bone phosphoproteins: role of glutamic acid-rich sequences in the nucleation of hydroxyapatite by bone sialoprotein. Biochem.J. 302 ( Pt 1):175-179


77


12. Tye CE, Hunter GK, Goldberg HA (2005) Identification of the type I collagen-binding domain of bone sialoprotein and characterization of the mechanism of interaction. J.Biol.Chem. 280:13487-13492


14. Anderson HC (1995) Molecular biology of matrix vesicles. Clin.Orthop.Relat Res.266-280 15. Reynolds JL, Joannides AJ, Skepper JN, McNair R, Schurgers LJ, Proudfoot D, Jahnen-

Dechent W, Weissberg PL, Shanahan CM (2004) Human vascular smooth muscle cells undergo vesicle-mediated calcification in response to changes in extracellular calcium and phosphate concentrations: a potential mechanism for accelerated vascular calcification in ESRD. J.Am.Soc.Nephrol. 15:2857-2867



18. Saoncella S, Echtermeyer F, Denhez F, Nowlen JK, Mosher DF, Robinson SD, Hynes RO, Goetinck PF (1999) Syndecan-4 signals cooperatively with integrins in a Rho-dependent manner in the assembly of focal adhesions and actin stress fibers. Proc.Natl.Acad.Sci.U.S.A 96:2805-2810

19. Bennett J, Scott KJ (1971) Quantitative staining of fraction I protein in polyacrylamide gels using Coomassie brillant blue. Anal.Biochem. 43:173-182

20. van Balkom BW, van Gestel RA, Brouwers JF, Krijgsveld J, Tielens AG, Heck AJ, van Hellemond JJ (2005) Mass spectrometric analysis of the Schistosoma mansoni tegumental sub-proteome. J.Proteome.Res. 4:958-966

21. Krijgsveld, J. et al. In-gel isoelectric focusing of peptdes as a tool for improved protein identification (2006) J Proteome Res . 5:1721-1730

22. Perkins DN, Pappin DJ, Creasy DM, Cottrell JS (1999) Probability-based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis 20:3551-3567

23. Kersey PJ, Duarte J, Williams A, Karavidopoulou Y, Birney E, Apweiler R (2004) The International Protein Index: an integrated database for proteomics experiments. Proteomics. 4:1985-1988

24. CSC/ICSM Proteomics Section Home Page. Available from: http://csc-fserve.hh.med.ic.ac.uk/msdb.html . 1-6-2004. Ref Type: Electronic Citation

25. Wu JT, Waterhouse WJ (1982) Identification of alpha-fetoprotein polymers. Artifacts of the isolation procedure. Clin.Chim.Acta 125:9-19

26. Wajih N, Borras T, Xue W, Hutson SM, Wallin R (2004) Processing and transport of matrix gamma-carboxyglutamic acid protein and bone morphogenetic protein-2 in cultured human vascular smooth muscle cells: evidence for an uptake mechanism for serum fetuin. J Biol.Chem. 279:43052-43060

27. Kawasaki T, Takahashi S, Ikeda K (1985) Hydroxyapatite high-performance liquid chromatography: column performance for proteins. Eur.J.Biochem. 152:361-371

28. Couchourel D, Escoffier C, Rohanizadeh R, Bohic S, Daculsi G, Fortun Y, Padrines M (1999) Effects of fibronectin on hydroxyapatite formation. J.Inorg.Biochem. 73:129-136

http://csc-fserve.hh.med.ic.ac.uk/msdb.html

http://csc-fserve.hh.med.ic.ac.uk/msdb.html

Chapter 3

78

29. Daculsi G, Pilet P, Cottrel M, Guicheux G (1999) Role of fibronectin during biological apatite crystal nucleation: ultrastructural characterization. J.Biomed.Mater.Res. 47:228-233

30. Mitsui K, Nakayama H, Akagi T, Nekooki M, Ohtawa K, Takio K, Hashikawa T, Nukina N (2002) Purification of polyglutamine aggregates and identification of elongation factor-1alpha and heat shock protein 84 as aggregate-interacting proteins. J.Neurosci. 22:9267-9277

31. Stefani M, Dobson CM (2003) Protein aggregation and aggregate toxicity: new insights into protein folding, misfolding diseases and biological evolution. J.Mol.Med. 81:678-699

32. Mitsui K, Nakayama H, Akagi T, Nekooki M, Ohtawa K, Takio K, Hashikawa T, Nukina N (2002) Purification of polyglutamine aggregates and identification of elongation factor-1alpha and heat shock protein 84 as aggregate-interacting proteins. J.Neurosci. 22:9267-9277

Chapter 4

The nitric oxide donor sodium nitroprusside inhibits mineralization in ATDC5 cells

Leonie F.A. Huitema1,2, Arie B. Vaandrager1, P. René van Weeren2, Ab Barneveld2, J. Bernd Helms1, Chris H.A. van de Lest1,2

1 Department of Biochemistry and Cell Biology, 2 Department of Equine Sciences, Faculty of Veterinary Medicine, and Graduate School Animal Health, Utrecht


Calcif. Tissue Int. 2006; 78; 171-178

Chapter 4

80

Abstract The aim of this study was to test whether the NO donor sodium nitroprusside has an effect on mineralization in ATDC5 cells. Mineralization in the ATDC5 cell culture was induced by addition of β-glycerophosphate or inorganic phosphate and visualized by staining precipitated calcium with an alizarin red stain and quantified using atomic absorption spectrometry. Sodium nitroprusside was shown to inhibit the mineralization of ADTC5 cells. This inhibition was not affected by inhibitors of guanylyl cyclase nor mimicked by a cGMP analog. Furthermore, sodium nitroprusside did not inhibit phosphate uptake nor inhibited apoptosis in the ATDC5 cells. These findings indicate that the NO donor sodium nitroprusside can specifically inhibit matrix mineralization via a cGMP-independent pathway and that the effect was not mediated by inhibition of phosphate transport or apoptosis. These results suggest a preventive role of NO in premature or pathological mineralization.

Sodium nitroprusside inhibits mineralization

81

Introduction Mineralization is an essential requirement for normal development and replacement of the cartilaginous skeleton via endochondral ossification [1]. Endochondral ossification is the process of cartilage formation, mineralization and replacement by bone that takes place at both metaphyseal and articular epiphyseal sites of the long bones [2]. This complex multistep process is tightly regulated and disturbances during the process of endochondral ossification may contribute to the pathogenesis of joint diseases [3]. Pathological mineralization can occur in all soft tissues [4]. For example, excessive mineral deposition accompanies osteoarthritis and calcification of cardiovascular tissues including arteries, heart valves, and cardiac muscle contributes to morbidity and mortality [3-7]. While pathological mineralization has long been considered to result from a mere physiochemical precipitation of calcium and phosphate, recent studies have provided evidence that vascular calcification is a regulated process, which has abundant similarities with bone formation [6-11]. This view is supported by the findings that vascular smooth muscle cells, like chondrocytes and osteoblasts, have the capacity to form matrix vesicles, which are believed to be the initiators of mineralization [12-16]. Furthermore, expression of a variety of bone-associated proteins has been found in atherosclerotic plaques [10, 11]. Although much effort has been devoted to identifying and characterizing the structure and/or components that initiate mineralization, the mechanisms regulating matrix mineralization, and the interplay between the various hormones involved in mineralization, are poorly understood. One of the agents implicated in bone growth and development as well as in the development of rheumatoid-, osteoarthritis and atherosclerosis is the gaseous substance nitric oxide (NO) [17-20]. NO is produced by various nitric oxide synthase (NOS) isoforms and may stimulate the soluble guanylyl cyclase (sGC) to generate the second messenger cGMP [21, 22]. Several studies have shown that cGMP is involved in bone formation [23] and that two known cGMP dependent protein kinases (cGK) isoforms (cGKI and cGKII) are present and distributed differently in the mouse growth plate chondrocytes. Mice deficient in the cGMP-dependent protein kinase II show dwarfism caused by a disturbed growth plate architecture [24]. In order to test a possible regulation of mineralization by NO generated by sodium nitroprusside (SNP), we used the chondrogenic mouse embryonal carcinoma-derived ATDC5 chondroprogenitor cells. ATDC5 cells cultured for 21 days in differentiation medium become hypertrophic in association with a dramatic elevation of alkaline phosphatases (ALP) activity. A subsequent medium switch to α-MEM has been shown to result in mineralization after another 14 days [25, 26]. In order to decrease the cell culture period required to induce mineralization, extra

Chapter 4

82

addition of phosphate (Pi) has been shown to increase the mineralization rate in the ATDC5 cell culture [27]. Materials and methods Materials Cell culture plastics were purchased from Greiner bio-one (Frickenhausen, Germany) and Nunc. Fetal bovine serum (FBS), antibiotics, and Insulin-Transferrin-Selenium-X (ITS-X) were obtained from Gibco. 1H-[1,2,4]Oxadiazolo[4,3-a]quinoxalin-1-one (ODQ) was provided by Calbiochem, cGMP analog (8-Br-cGMP) was obtained from BioLog Life Science Institute (Bremen, Germany) and α-minimum essential medium (α-MEM) was from Biochrom AG (Berlin, Germany). [32P] orthophosphate was purchased from Amersham and the in situ cell death detection kit was obtained from Roche Applied Science. All other chemicals were obtained from Sigma. Cell line and culture conditions ATDC5 cells (Riken Cell Bank, Tsukuba, Japan) were cultured and maintained in a 1:1 mixture of Dulbecco’s modified Eagle medium and Ham’s F12 medium (DMEM/F12) containing 5% FBS, 10 µg/ml human transferrin, 30 nM sodium selenite, 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2 – 95% air. Cells were subcultured once a week using trypsin/EDTA solution. For induction of chondrogenesis, ATDC5 cells were cultured in DMEM/F12 containing 5% FBS, ITS-X, 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2 – 95% air. Medium was replaced every 2 days. Mineralization was induced by transferring the cells to α-MEM containing 5% FBS, ITS-X, 2 or 5 mM phosphate, 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2 – 95% air. Cells were seeded with an initial density of 15,000 cells/cm2 and medium was replaced every 2-3 days. For alizarin red staining, deposited calcium quantification and Pi transport cells were cultured in 4 wells plates (1.9 cm2/well). For cell imaging cells were cultured in Lab-Tek chamber slides (16 x 8 wells, 0.8 cm2/well, permanox). Alizarin red staining Mineralization was visualized by the staining of precipitated calcium with alizarin red. Cell layers were washed twice with PBS and fixed for 15 minutes in 70% ethanol, followed by incubation of 500 µl 0.5% alizarin red solution for 5 minutes. Excessive dye was removed by washing the cells 4 times with distilled H2O.


83

Alizarin red staining was examined visually, and photographed with a Nikon coolpix 9500 digital camera. Measurement of total calcium content The total amount of calcium present in ATDC5 cell layers was measured by atomic absorption spectroscopy at 422.7 nm (Perkin Elmer 3300, Perkin Elmer Corp., Norwalk, Cl, USA). Therefore, cells were washed twice with PBS, hydrolyzed in 400 µl 6 M HCL and diluted twice with 0.5% lanthanum citrate solution to control interferences with phosphate. Measurement of phosphate transport ATDC5 cell layers were incubated for 5 minutes in 1 ml transport buffer (20 mM HEPES, 135 mM NaCl, 2 mM CaCl2, 2 mM MgCl2, 5 mM KCl and 100 µM NaH2PO4/Na2HPO4, pH 7.4) at 37°C. Pi transport was then determined by adding an additional 0.5 ml HEPES buffer containing 2 µCi/ml [32P] orthophosphate. After 20 minutes extracellular 32Pi was removed by washing the cell layers 3 times with ice-cold PBS. The cells were then dissolved in 0.2 M NaOH and 32Pi was quantified using a standard liquid scintillation technique. The sodium independent components of Pi uptake were determined by replacing NaCl with choline chloride. Detection of apoptosis Apoptosis of the ATDC5 cells was determined by the in situ cell death detection kit TMR red (Roche), based on labelling DNA strand breaks (TUNEL technology). Cells were fixed in 4% formaldehyde in PBS overnight at 4°C, free aldehyde groups were blocked with 50 mM glycine-PBS (pH 7.3) for 30 minutes, after which the cells were permeabilized with 0.1% Triton X-100 in 0.1% sodium citrate on ice for 2 minutes. Labeling was performed according to the manufacturer’s protocol. All nuclei in the fixed cell population were visualized by staining with 2 µg/ml Hoechst 33342. To quantify the amount of apoptotic cells in a whole cell population, Hoechst labeled nuclei and apoptotic nuclei from three independent experiments were counted in five different microscopic fields per experiment. The total amount of Hoechst labeled nuclei in a field was taken as hundred percent. The samples were imaged using a digital imaging microscope, Leica DMRA fluorescence microscope with additional photomatrix coolsnap CCD camera. Statistical analysis Experiments were performed at least three times in duplo. Statistical analyses were performed using an analysis of variance (ANOVA), followed by Dunnett's Multiple Comparison Test (one-way ANOVA), and in case of a two-way ANOVA

Chapter 4

84

Bonferoni’s post-hoc test was performed. Values of P<0.05 where considered to be significant. Values where expressed as mean ± SEM.

Figure 1. Calcium deposition in the culture of ATDC5 cells. Cells were cultured for 14 days in differentiation medium and subsequently in (A) mineralization medium + 2 mM Pi for 48 hours, (B) mineralization medium + 2 mM Pi + 100 µM SNP for 48 hours, (C) mineralization medium for 48 hours, (D) mineralization medium + 5 mM Pi for 24 hours, (E) ) mineralization medium + 5 mM Pi + 100 µM SNP for 24 hours, (F) mineralization medium for 24 hours, (G) mineralization medium + 10 mM β-GP for 48 hours, (H) mineralization medium + 10 mM β-GP + 100 µM SNP for 48 hours and (I) mineralization medium for 48 hours. Cell cultures were fixed in 70% ethanol and stained with 0.5% Alizarin red. The red color indicates mineralization spots.


85

Results Effect of sodium nitroprusside on mineralization In our search for agents that affect matrix mineralization we tested the NO donor SNP. As shown in Fig. 1 addition of 2 mM Pi to the mineralization medium resulted in mineralization, which was inhibited by addition of SNP (100 µM) to the mineralization medium. SNP (100 µM) also inhibited mineralization when 5 mM Pi (incubation time 24 hours) was added to the mineralization medium. Supplementation with β-glycerophosphate (β-GP) has also been shown to facilitate mineralization by numerous investigators [28-33]. Therefore, it was determined whether SNP could inhibit mineralization induced by β-GP. Fig.1 shows that addition of 10 mM β-GP to the mineralization medium induced mineralization in the ATDC5 cell culture, which was inhibited by additional SNP (100 µM) to the mineralization medium. Next, the dose-response effect of SNP on mineralization (5 mM Pi for 24 hours) was investigated by quantification of the total calcium content by atomic absorption spectrometry. As shown in Fig. 2, SNP at a concentration of 10 µM already inhibited mineralization by more than 60%.

Figure 2. Dose-response effect of SNP on mineralization in the ATDC5 cell culture. Cells were grown 14 days in differentiation medium and subsequently 24 hours in mineralization medium with addition of 5 mM Pi and different concentrations SNP. Calcium was extracted and measured by atomic absorption spectrometry as described in the Materials and Methods section. Results are expressed as mean ± SEM (n=6). * P< 0.05, **P< 0.01, ***P<0.001 with respect to the corresponding control value.

Chapter 4

86

Involvement of cGMP in effect of sodium nitroprusside Because NO produced by SNP could activate sGC, an enzyme known to produce cGMP, it was examined whether SNP inhibits mineralization via the cGMP-signaling pathway. In this case, inhibition of sGC should reverse the inhibiting effect on mineralization. Fig. 3 shows that the sGC inhibitor ODQ (100 µM) could not reverse the inhibiting effect of SNP on mineralization. Similar results were obtained with another sGC inhibitor (LY83583) (data not shown). Furthermore, a cell permeable cGMP analog, 8-bromo-cGMP (50 µM), was unable to mimic the effect of SNP, as it did not inhibit the Pi-induced mineralization in the ATDC5 cell culture (Fig. 3).

Figure 3. Test whether SNP effect is cGMP-dependent. Cells were cultured in differentiation medium for 14 days and subsequently 24 hours in mineralization medium + 5 mM Pi with addition of 100 µM SNP or 100 µM SNP + 100 µM ODQ (sGC inhibitor) or 50 µM 8-bromo-cGMP (cGMP analog). Calcium was extracted and measured by atomic absorption spectrometry as described in the Materials and Methods section Results are expressed as mean ± SEM (n=6). ** P< 0.01 with respect to the control value. Pi transport studies It has been shown that inorganic phosphate (Pi) can induce apoptosis and mineralization in chondrocytes [34, 35]. Pi transport into chondrocytes is mediated by the sodium-dependent Pi type III transporter Glvr-1, which has been demonstrated to be expressed in a confined population of early hypertrophic chondrocytes in vivo [36]. To further explore the mechanism by which SNP


87

inhibits mineralization, we tested whether SNP affects phosphate transport. Therefore, ATDC5 cells were cultured for 14 days in differentiation medium and subsequently for 24 hours in either the same differentiating medium with or without addition of 100 µM SNP or in α-MEM with or without addition of 100 µM SNP. Phosphate transport was measured by incubation with 32Pi for 20 min as described in the Materials & Methods section. As a negative control, ATDC5 cells were incubated in medium without sodium resulting in a lack of Pi transport (Fig. 4). As shown in Fig. 4, 24h pre-incubation with 100 µM SNP did not decrease Pi uptake in ATDC5 cells. Of note, ATDC5 cells appear to take up more Pi when cultured 24 hours in α-MEM, but mineralization under these conditions could not be detected (data not shown).

Figure 4. Pi transport in ATDC5 cells. Cells were cultured 14 days in differentiation medium and subsequently 24 hours in either differentiation medium (open bars) or in mineralization medium (solid bars) in presence or absence of SNP and with or without sodium. Pi transport activity was determined as described in the Material and Methods section. Results are expressed as mean ± SEM (n=6). Cells cultured 24 hours in mineralization medium exhibit significantly more Pi transport than cells cultured in differentiation medium (*P<0.05).

Chapter 4

88

Figure 5. DNA fragmentation (TUNEL) and nucleus (Hoechst) stain in ATDC5 cells. Cells were cultured 14 days in differentiation medium and subsequently 24 hours in (A) differentiation medium, (B) mineralization medium + 5 mM Pi (C) mineralization medium + 5 mM Pi + 100 µM SNP, or (D) mineralization medium + 5 mM Pi + 2 µM staurosporin, or (E) mineralization medium + 5 mMPi + 100 µM SNP + 2 µM staurosporin. Cell layers were fixed in 4% paraformaldehyde. Subsequently, a TUNEL reaction and Hoechst stain was performed as described in the Materials and Methods section and was visualized by a digital imaging microscope.


89

Analysis of apoptosis Pi has been demonstrated to induce apoptosis related mineralization in chondrogenic cells [27, 34, 35, 37], whereas low concentrations of NO have been shown to inhibit cell death in other cell types [38-41]. Therefore, we examined whether SNP inhibits mineralization by inhibiting cell death. Apoptotic cell death was assessed by an immunohistochemical in situ cell death detection kit based on labeling of DNA strand breaks (TUNEL technology). As shown in Fig. 5, no differences in apoptosis were observed between cells cultured in differentiation medium (16.7 ± 4.8%), mineralization medium (18.1 ± 2.5%) or in mineralization medium with SNP (18.7 ± 5%). To test whether or not apoptosis is involved in mineralization, we induced apoptosis in the ATDC5 cell culture by addition of 2 µM staurosporin to the mineralization medium [42]. Staurosporin increased the mineralization rate (Fig. 6) and induced massive cell death (80 ± 15%) in the ATDC5 cell culture (Fig. 5). Under these conditions SNP did not inhibit apoptosis, but it inhibited mineralization to the same extent both in the absence and presence of staurosporin (Fig. 6). Furthermore, we observed massive cell death in ATDC5 cells treated with 1 mM SNP alone (data not shown) but almost no mineralization was observed under these conditions (Fig. 2). Together, these results show that SNP does not inhibit mineralization by inhibiting apoptosis.

Figure 6. Effects of staurosporin and SNP on calcium deposition in ATDC5 cells. Cells were cultured 14 days in differentiation medium, and subsequently 24 hours in (A) mineralization medium + 5 mM Pi, (B) mineralization medium + 5 mM Pi + 2 µM staurosporin, or (C) mineralization medium + 5 mM Pi + 2 µM staurosporin + 100 µM SNP. Cell layers were fixed in 70% ethanol and stained with 0.5% Alizarin red. Discussion In our search for agents that can affect mineralization we found that the NO donor SNP inhibits mineralization by more than 60% at a concentration of 10 µM. NO

Chapter 4

90

generated by SNP, is a diffusible, multifunctional, transcellular messenger that has been implicated in many physiological and pathological conditions [19, 20, 22, 38-41]. NO has been shown to activate the cGMP-signaling pathway via activation of sGC [22]. Our data show that the NO donor SNP inhibits mineralization via a cGMP independent mechanism, as its effect was not affected by an inhibitor of sGC and could not be mimicked by a cGMP analog. In vivo and in vitro results have shown that Pi (together with Ca), induces mineralization which is accompanied by apoptosis, suggesting a role for both ions and apoptosis during endochondral ossification [15, 16, 33, 34, 42]. NO has been shown to inhibit caspases by S-nitrosylation, resulting in an inhibition of apoptosis [43, 44]. Therefore, we tested whether NO generated by 100 µM SNP inhibited mineralization by inhibiting apoptosis. However, our TUNEL data showed no difference in apoptosis in the presence or absence of a low dose of SNP. In contrast to other studies, who found that Pi induced apoptosis of differentiated chondrocytes [27, 34, 35, 37, 42], we observed no increase in apoptosis after the mineralization medium switch including addition of 4 mM Pi. A possible explanation for the difference in results with that observed in previous studies is that we performed all our experiments at early stages of mineralization. Induction of apoptosis with staurosporin increased cell death and mineralization in our cell culture system, showing an association between apoptosis and mineralization. The staurosporin induced cell death was also observed in cultures treated with SNP, but here mineralization was inhibited. High concentrations of NO generated by 1 mM SNP have been shown to induce apoptosis in cultured chondrocytes [45], and apoptotic bodies or matrix vesicles have been isolated from these cultures and were shown to have calcifying capacities [45, 46]. Our results show that millimolar concentrations of SNP can also inhibit mineralization, despite the induction of apoptosis. We therefore speculate that SNP may inhibit the nucleation of calcium phosphate in apoptotic bodies, possibly by interacting with mineralization stimulating factors present in these vesicles. Alternatively, SNP may inhibit the influx of Ca or Pi into the vesicles. The inhibition of the calcium channel activity of annexin V was reported to prevent the mineralization of matrix vesicles [47, 48]. Inhibiting Pi uptake has been shown to inhibit mineralization [27]. Our data show that phosphate transport is not inhibited by NO generated by SNP in whole cells, but we cannot exclude that SNP affects phosphate transport in vesicles. In bone, NO has been shown to have a role in the pathogenesis of osteoarthritis and rheumatoid arthritis as well as a physiological role in bone growth and development [20]. NO can be produced by a family of NO synthase iso-enzymes (NOS), endothelial NOS (eNOS), neuronal NOS (nNOS) and inducible NOS (iNOS). Temporal expression of all three isoforms of NOS during rat femoral fracture healing has been demonstrated. Suppression of NOS with NOS inhibitors


91

impairs fracture healing, which can be reversed with supplementation of a NO donor [49, 50]. During the initial phase of fracture healing high concentrations of NO were generated by iNOS. At the stage of endochondral ossification, iNOS declined and was replaced by low NO producing eNOS and nNOS. This indicates that differential expression NOS regulates fracture repair, either enhancing it or to suppress unwanted bone formation. iNOS expression has also been associated with apoptosis, the stimulation of matrix metallo-proteinases, and suppression of proteoglycan synthesis of chondrocytes, resulting in destruction of cartilage [19, 20]. Expression of iNOS may also cause inhibition of mineralization, since lipopolysacharide and interleukin-1, both strong inducers of iNOS, were shown to inhibit mineralization in cartilage organ cultures [51]. Considering the inhibitory effect of the NO donor SNP on mineralization found in this study, we may speculate that the NO produced by iNOS may be generated preferentially in the undifferentiated and early hypertrophic chondrocytes in order to prevent premature mineralization in articular cartilage. NO may also be effective in preventing the pathological calcium crystal formation observed in non-chondrogenic tissues such as vasculature. High Pi levels as occurring during renal dysfunction are known to be a risk factor for developing calcification of atherosclerotic plaques [52]. Interestingly, the NO donor SNP is used therapeutically in cardiovascular diseases for potent vasodilating effects. It is therefore tempting to speculate that the beneficial effects of this drug may also include an inhibition of calcification of the vasculature. Acknowledgement We thank Hugo S. Wouterse for expert assistance with the atomic absorption spectroscopy. We also thank the Centre for Cell Imaging at the Faculty of Veterinary Medicine, Utrecht University for using the fluorescence microscope. Reference List 1 Kronenberg HM 2003 Developmental regulation of the growth plate. Nature 423:332-336. 2 Cancedda R, Castagnola P, Cancedda FD, Dozin B, Quarto 2000 Developmental control of

chondrogenesis and osteogenesis. Int J Dev Biol 44:707-714. 3 Shum L, Coleman CM, Hatakeyama Y, Tuan RS 2003 Morphogenesis and

dysmorphogenesis of the appendicular skeleton. Birth Defects Res C Embryo Today 69:102-122.

4 Giachelli CM 1999 Ectopic calcification: gathering hard facts about soft tissue mineralization. Am J Pathol 154:671-675.

Chapter 4

92

5 Magne D, Julien M, Vinatier C, Merhi-Soussi F, Weiss P, Guicheux J 2005 Cartilage formation in growth plate and arteries: from physiology to pathology. Bioessays 27:708-716.

6 Kalya S, Rosenthal AK 2005 Extracellular matrix changes regulate calcium crystal formation in articular cartilage. Curr Opin Rheumatol 17:325-329.

7 Demer LL, Tintut Y 2003 Mineral exploration: search for the mechanism of vascular calcification and beyond: the 2003 Jeffrey M. Hoeg Award lecture. Arterioscler Thromb Vasc Biol 23:1739-1743.

8 Shanahan CM, Cary NR, Salisbury JR, Proudfoot D, Weissberg PL, Edmonds ME 1999 Medial localization of mineralization-regulating proteins in association with Monckeberg's sclerosis: evidence for smooth muscle cell-mediated vascular calcification. Circulation 100:2168-2176.

9 Bostrom K, Watson KE, Horn S, Wortham C, Herman IM, Demer LL 1993 Bone morphogenetic protein expression in human atherosclerotic lesions. J Clin Invest 91:1800-1809.

10 Bobryshev YV 2005 Transdifferentiation of smooth muscle cells into chondrocytes in atherosclerotic arteries in situ: implications for diffuse intimal calcification. J Pathol 205:641-650.

11 Tyson KL, Reynolds JL, McNair R, Zhang Q, Weissberg PL, Shanahan CM 2003 Osteo/chondrocytic transcription factors and their target genes exhibit distinct patterns of expression in human arterial calcification. Arterioscler Thromb Vasc Biol 23:489-494.

12 Hsu HH, Camacho NP 1999 Isolation of calcifiable vesicles from human atherosclerotic aortas. Atherosclerosis 143:353-362.

13 Hsu HH, Camacho NP, Sun F, Tawfik O, Aono H 2000 Isolation of calcifiable vesicles from aortas of rabbits fed with high cholesterol diets. Atherosclerosis 153:337-348.

14 Hsu HH, Camacho NC, Tawfik O, Sun F 2002 Induction of calcification in rabbit aortas by high cholesterol diets: roles of calcifiable vesicles in dystrophic calcification. Atherosclerosis 161:85-94.

15 Kirsch T, Nah HD, Shapiro IM, Pacifici M 1997 Regulated production of mineralization-competent matrix vesicles in hypertrophic chondrocytes. J Cell Biol 137:1149-1160.

16 Anderson HC 1995 Molecular biology of matrix vesicles. Clin Orthop 266-280. 17 Li H, Forstermann U 2000 Nitric oxide in the pathogenesis of vascular disease. J Pathol

190:244-254. 18 Alexander RW, Dzau VJ 2000 Vascular biology: the past 50 years. Circulation 102:IV112-

IV116. 19 Jang D, Murrell GA 1998 Nitric oxide in arthritis. Free Radic Biol Med 24:1511-1519. 20 van't Hof RJ, Ralston SH 2001 Nitric oxide and bone. Immunology 103:255-261. 21 Fujishige K, Kotera J, Yanaka N, Akatsuka H, Omori K 1999 Alteration of cGMP

metabolism during chondrogenic differentiation of chondroprogenitor-like EC cells, ATDC5. Biochim Biophys Acta 1452:219-227.

22 Krumenacker JS, Hanafy KA, Murad F 2004 Regulation of nitric oxide and soluble guanylyl cyclase. Brain Res Bull 62:505-515.

23 Ozasa A, Komatsu Y, Yasoda A, Miura M, Sakuma Y, Nakatsuru Y, Arai H, Itoh N, Nakao K 2005 Complementary antagonistic actions between C-type natriuretic peptide and the MAPK pathway through FGFR-3 in ATDC5 cells. Bone 36:1056-1064.

24 Pfeifer A, Aszodi A, Seidler U, Ruth P, Hofmann F, Fassler R 1996 Intestinal secretory defects and dwarfism in mice lacking cGMP-dependent protein kinase II. Science 274:2082-2086. 25 Shukunami C, Shigeno C, Atsumi T, Ishizeki K, Suzuki F, Hiraki Y 1996 Chondrogenic

differentiation of clonal mouse embryonic cell line ATDC5 in vitro: differentiation-


93

dependent gene expression of parathyroid hormone (PTH)/PTH-related peptide receptor. J Cell Biol 133:457-468.

26 Shukunami C, Ishizeki K, Atsumi T, Ohta Y, Suzuki F, Hiraki Y 1997 Cellular hypertrophy and calcification of embryonal carcinoma-derived chondrogenic cell line ATDC5 in vitro. J Bone Miner Res 12:1174-1188.

27 Magne D, Bluteau G, Faucheux C, Palmer G, Vignes-Colombeix C, Pilet P, Rouillon T, Caverzasio J, Weiss P, Daculsi G, Guicheux J 2003 Phosphate is a specific signal for ATDC5 chondrocyte maturation and apoptosis-associated mineralization: possible implication of apoptosis in the regulation of endochondral ossification. J Bone Miner Res 18:1430-1442.

28 Bellows CG, Aubin JE, Heersche JN 1991 Initiation and progression of mineralization of bone nodules formed in vitro: the role of alkaline phosphatase and organic phosphate. Bone Miner 14:27-40.

29 Johnson KA, Hessle L, Vaingankar S, Wennberg C, Mauro S, Narisawa S, Goding JW, Sano K, Millan JL, Terkeltaub R 2000 Osteoblast tissue-nonspecific alkaline phosphatase antagonizes and regulates PC-1. Am J Physiol Regul Integr Comp Physiol 279:R1365-R1377.

30 Kergosien N, Sautier J, Forest N 1998 Gene and protein expression during differentiation and matrix mineralization in a chondrocyte cell culture system. Calcif Tissue Int 62:114-121.

31 Loty S, Sautier JM, Loty C, Boulekbache H, Kokubo T, Forest N 1998 Cartilage formation by fetal rat chondrocytes cultured in alginate beads: a proposed model for investigating tissue-biomaterial interactions. J Biomed Mater Res 42:213-222.

32 Lunstrum GP, Keene DR, Weksler NB, Cho YJ, Cornwall M, Horton WA 1999 Chondrocyte differentiation in a rat mesenchymal cell line. J Histochem Cytochem 47:1-6.

33 Stanford CM, Jacobson PA, Eanes ED, Lembke LA, Midura RJ 1995 Rapidly forming apatitic mineral in an osteoblastic cell line (UMR 106-01 BSP). J Biol Chem 270:9420-9428.

34 Mansfield K, Rajpurohit R, Shapiro IM 1999 Extracellular phosphate ions cause apoptosis of terminally differentiated epiphyseal chondrocytes. J Cell Physiol 179:276-286.

35 Mansfield K, Teixeira CC, Adams CS, Shapiro IM 2001 Phosphate ions mediate chondrocyte apoptosis through a plasma membrane transporter mechanism. Bone 28:1-8.

36 Palmer G, Zhao J, Bonjour J, Hofstetter W, Caverzasio J 1999 In vivo expression of transcripts encoding the Glvr-1 phosphate transporter/retrovirus receptor during bone development. Bone 24:1-7.

37 Mansfield K, Pucci B, Adams CS, Shapiro IM 2003 Induction of apoptosis in skeletal tissues: phosphate-mediated chick chondrocyte apoptosis is calcium dependent. Calcif Tissue Int 73:161-172.

38 Fiscus RR 2002 Involvement of cyclic GMP and protein kinase G in the regulation of apoptosis and survival in neural cells. Neurosignals 11:175-190.

39 Zech B, Kohl R, von Knethen A, Brune B 2003 Nitric oxide donors inhibit formation of the Apaf-1/caspase-9 apoptosome and activation of caspases. Biochem J 371:1055-1064.

40 Chan SL, Fiscus RR 2003 Guanylyl cyclase inhibitors NS2028 and ODQ and protein kinase G (PKG) inhibitor KT5823 trigger apoptotic DNA fragmentation in immortalized uterine epithelial cells: anti-apoptotic effects of basal cGMP/PKG. Mol Hum Reprod 9:775-783.

41 Yoshioka Y, Yamamuro A, Maeda S 2003 Nitric oxide at a low concentration protects murine macrophage RAW264 cells against nitric oxide-induced death via cGMP signaling pathway. Br J Pharmacol 139:28-34.

Chapter 4

94

42 Kirsch T, Wang W, Pfander D 2003 Functional differences between growth plate apoptotic bodies and matrix vesicles. J Bone Miner Res 18:1872-1881

43 Matsumoto A, Comatas KE, Liu L, Stamler JS 2003 Screening for nitric oxide-dependent protein-protein interactions. Science 301:657-661.

44 Mahidhara RS, Hoffman RA, Huang S, Wolf-Johnston A, Vodovotz Y, Simmons RL, Billiar TR 2003 Nitric oxide-mediated inhibition of caspase-dependent T lymphocyte proliferation. J Leukoc Biol 74:403-411.

45 Hashimoto S, Ochs RL, Rosen F, Quach J, McCabe G, Solan J, Seegmiller JE, Terkeltaub R, Lotz M 1998 Chondrocyte-derived apoptotic bodies and calcification of articular cartilage. Proc Natl Acad Sci U S A 95:3094-3099.

46 Cheung HS, Ryan LM 1999 Phosphocitrate blocks nitric oxide-induced calcification of cartilage and chondrocyte-derived apoptotic bodies. Osteoarthritis Cartilage 7:409-412.

47 Kirsch T, Harrison G, Golub EE, Nah HD: The roles of annexins and types II and X collagen in matrix vesicle-mediated mineralization of growth plate cartilage. J Biol Chem 2000;275:35577-35583.

48 Wang W, Xu J, Kirsch T 2003 Annexin-mediated Ca2+ influx regulates growth plate chondrocyte maturation and apoptosis. J Biol Chem 278:3762-3769.

49 Hauselmann HJ, Stefanovic-Racic M, Michel BA, Evans CH 1998 Differences in nitric oxide production by superficial and deep human articular chondrocytes: implications for proteoglycan turnover in inflammatory joint diseases. J Immunol 160:1444-1448.

50 Hayashi T, Abe E, Yamate T, Taguchi Y, Jasin HE 1997 Nitric oxide production by superficial and deep articular chondrocytes. Arthritis Rheum 40:261-269.

51 Scheller M, Zimmermann B, Bernimoulin JP, Scholz P 1995 Induction of metalloproteinase activity, cartilage matrix degradation and inhibition of endochondral mineralization in vitro by E. coli lipopolysaccharide is mediated by interleukin 1 alpha. Cytokine 7:331-337.

52 Vattikuti R, Towler DA 2004 Osteogenic regulation of vascular calcification: an early perspective. Am J Physiol Endocrinol Metab 286:E686-E696.

Chapter 5

Iron ions derived from the nitric oxide donor sodium nitroprusside inhibit mineralization.

Leonie F.A. Huitema1,2, P. René van Weeren2, Ab Barneveld2, Chris H.A. van de Lest1,2, J. Bernd Helms1, Arie B. Vaandrager1

1 Department of Biochemistry and Cell Biology, 2 Department of Equine Sciences,

Faculty of Veterinary Medicine, and Graduate School Animal Health, Utrecht University, Utrecht, the Netherlands

Eur. J. Pharm. 2006; 542;48-53

Chapter 5

96

Abstract Sodium nitroprusside (SNP) is a nitric oxide (NO) donor drug, which is therapeutically used as a vasodilating drug in heart transplantations. In our previous study it was found that SNP at a concentration of 100 µM inhibited mineralization in a cell culture system, indicating that the beneficial effects of this drug may also include inhibition of vascular calcification. The aim of this study was to investigate which bioactive compounds generated from SNP inhibit mineralization. ATDC5 cells were grown for 14 days and mineralization was induced by addition of 5 mM phosphate for 24 h. Mineralization was determined by staining precipitated calcium with an alizarin red stain. It was found that the NO donors S-nitrosogluthatione and S-nitroso-N-acetylpencillamine were not able to inhibit mineralization and NO scavengers could not antagonize the inhibiting effect of SNP on mineralization. The iron chelator deferoxamine (200 µM) antagonized the inhibiting effect on mineralization mediated by SNP and ammonium iron sulfate inhibited mineralization in a dose-dependent manner (10-100 µM). Furthermore, iron ions (30 µM) were detected to be released from SNP in the cell culture. These data show that the iron moiety of sodium nitroprusside, rather than nitric oxide inhibits mineralization.

Iron ions from SNP inhibit mineralization

97

Introduction Mineralization is essential for normal skeletal development, which is mainly accomplished through the function of two cell types, osteoblasts and chondrocytes [1-3]. In contrast, soft tissues do not mineralize under normal conditions [4]. However under certain pathological conditions some organs in particular blood vessels, are prone to mineralization. For example, atherosclerosis, aging, diabetes and end-stage renal disease predispose arteries to calcification, which is highly associated with mortality in patients with cardiovascular disease [5]. Growing evidence suggests that vascular calcification, like bone formation, is a highly regulated process, involving both inducers and inhibitors [5]. This view is supported by the findings that vascular smooth muscle cells, like chondrocytes and osteoblasts, have the capacity to form matrix vesicles, which are believed to be the initiators of mineralization [6-10]. Furthermore, bone-related non-collagenous proteins have been found in atherosclerotic plaques [11;12]. Based on these findings it is likely that cell mediated processes maintain a balance between induction and inhibition of mineralization in the artery, such that mineralization is normally prevented. In our previous study it was found that the nitric oxide (NO) donor drug sodium nitroprusside (SNP) [Na2(Fe(CN)5NO)] inhibits mineralization in a cell culture system using the ATDC5 cell line [13]. SNP is a widely used NO donor drug for in vitro and in vivo experimental studies [14;15]. The major target of NO is the heme iron of soluble guanylyl cyclase, resulting in activation of the enzyme. This results in an increase in intracellular levels of cyclic guanosine-3’,5’monophosphate (cGMP), which acts as a second messenger [16]. However, our results showed that the inhibiting effect of mineralization generated by SNP was cGMP-independent. Although SNP is generally described in biological and medical literature as an NO donor drug, it has also been shown to decompose to other bioactive compounds, such as cyanide or pentacyanoferrate complex (Fig. 1) [15;17;18].

Figure 1. Structure of sodium nitroprusside.

Chapter 5

98

Furthermore, SNP may undergo redox cycling with the production of superoxide anion that in turn produces hydrogen peroxide [18]. It has also been shown that the iron moiety of SNP may mediate pro-oxidative properties of SNP, such as hydroxyl radical generation, lipid peroxidation and cytotoxity [19]. Therefore, the aim of the present study was to elucidate which bioactive products generated by SNP inhibit mineralization. Materials and methods Materials Cell culture plastics were purchased from Greiner bio-one (Frickenhausen, Germany) and Nunc. Fetal bovine serum (FBS), antibiotics, and Insulin-Transferrin-Selenium-X (ITS-X) were obtained from Gibco BRL (Grand Island, NY). All other chemicals were obtained from Sigma (Poole, UK). Cell line and culture conditions ATDC5 cells (Riken Cell Bank, Tsukuba, Japan) were cultured and maintained in a 1:1 mixture of Dulbecco’s modified Eagle medium and Ham’s F12 medium (DMEM/F12) containing 5% FBS, 10 µg/ml human transferrin, 30 nM sodium selenite, 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2 – 95% air. Cells were subcultured once a week using trypsin/EDTA solution. For induction of differentiation, ATDC5 cells were cultured 14 days in DMEM/F12 containing 5% FBS, ITS-X, 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2 – 95% air. Medium was replaced every 2 days. Mineralization was induced by transferring the cells to α-MEM containing 5% FBS, ITS-X, 5 mM inorganic phosphate (Pi), 100 IU/ml penicillin, and 100 µg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2 – 95% air for 24 h. Cells were seeded with an initial density of 15,000 cells/cm2 in 4 wells plates (1.9 cm2/well). Medium was replaced every 2-3 days. Alizarin red staining Mineralization was visualized by the staining of precipitated calcium with alizarin red. Cell layers were washed twice with PBS and fixed for 15 min in 70% ethanol, followed by incubation of 500 µl 0.5% alizarin red solution for 5 min. Excessive dye was removed by washing the cells 4 times with distilled H2O. To quantify mineralization in the ATDC5 cell culture the alizarin red stained cultures were incubated with 100 mM cetylpyridinium chloride for 10 min to release calcium bound alizarin red into solution. The absorbance of the released alizarin red stain


99

was measured at 595 nm using a Bio-rad microplate reader (Benchmark). Data are expressed as arbitrary units alizarin red stain. Iron determination Iron was determined using Ferrozine. To 200 µl cell cultured medium 40 µl 10 mM Ferrozine in 50 mM HEPES, pH 7, was added. After allowing for color development (2 h), the absorbance of the Ferrozine assay mixture was determined at 525 nm using a Bio-rad microplate reader (Benchmark). Statistical analysis Experiments were performed at least three times in duplo. Statistical analyses were performed using an analysis of variance (ANOVA), followed by Dunnett's Multiple Comparison Test. Values of P<0.01 where considered to be significant. Values were expressed as mean ± S.E.M. Results Effect of NO donors and NO scavengers on mineralization In our previous study it was found that SNP at a concentration of 100 µM inhibited mineralization [13]. To investigate whether the inhibiting effect on mineralization by SNP (100 µM) was effectuated by nitric oxide, different NO donors were tested.

Figure 2. Effect of NO donors on mineralization. ATDC5 cells were grown 14 days in differentiation medium and subsequently 24 h in mineralization medium (w/o Pi), or with addition of 5 mM Pi (M), or 5 mM Pi + 100 µM sodium nitroprusside (SNP), or 5 mM Pi + 100 µM S-nitroso-N-acetylpencillamine (SNAP), or 5mM Pi + 100 µM S-nitrosogluthatione (GSNO). Cell cultures were fixed in 70% ethanol and stained with 0.5% Alizarin red. Subsequently the alizarin red stain was extracted and measured by a spectrophotometer as described in the Materials and Methods section. Results are expressed as mean ± S.E.M., * P< 0.01 with respect to the corresponding control value (M).

Chapter 5

100

As shown in Fig. 2 the NO donor S-nitroso-N-acetylpenicillamine (SNAP) and the NO donor S-nitrosogluthatione (GSNO) did not inhibit mineralization in ATDC5 cells at concentrations of 100 µM. Higher concentrations up to 1 mM of SNAP and GSNO did also not prevent mineralization (data not shown). Next, it was investigated whether scavenging NO could antagonize the action of SNP. Therefore, two different NO scavengers were tested, 2-Phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl 3-oxide (PTIO) and hemoglobin (Hb). Fig. 3 shows that both NO scavengers could not reverse the inhibiting effect of SNP on mineralization. These results suggest that the inhibiting effect of SNP cannot be ascribed to the release of NO.

Figure 3. The NO scavengers PTIO and Hb could not prevent the inhibiting effect of SNP on mineralization. ATDC5 cells were grown 14 days in differentiation medium and subsequently 24 h in mineralization medium with addition of 5 mM Pi (M), or 5 mM Pi + 100 µM SNP, or 5 mM Pi + 100 µM SNP + 200 µM 2-Phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl 3-oxide (PTIO), or 5 mM Pi + 100 µM SNP + 2 mg/ml hemoglobin (Hb). Cell cultures were fixed in 70% ethanol and stained with 0.5% Alizarin red. Subsequently the alizarin red stain was extracted and measured by a spectrophotometer as described in the Materials and Methods section. Results are expressed as mean ± S.E.M., * P< 0.01 with respect to the control value (M). Effect of (iron) cyanide on mineralization Cyanide is a toxic compound that can be released from SNP (Fig. 1). Therefore, the effect of (iron) cyanide on mineralization was investigated. As shown in Fig. 4, 100 µM potassium cyanide, potassium ferrocyanide and potassium ferricyanide could not inhibit mineralization, excluding the possibility that the inhibiting effect of mineralization may be ascribed to cyanide or the ferricyanide complex.


101

Figure 4. Effect of (iron) cyanide on mineralization. ATDC5 cells were grown 14 days in differentiation medium and subsequently 24 h in mineralization medium with addition of 5 mM Pi (M), or 5 mM Pi + 100 µM sodium nitroprusside (SNP), or 5 mM Pi + 100 µM potassium ferrocyanide (FeIICN), or 5mM Pi + 100 µM potassium ferricyanide (FeIIICN), or 5mM Pi + 100 µM potassium cyanide (KCN). Cell cultures were fixed in 70% ethanol and stained with 0.5% Alizarin red. Subsequently the alizarin red stain was extracted and measured by a spectrophotometer as described in the Materials and Methods section. The results represent the means ± S.E.M. of three independent experiments each performed in duplo, * P< 0.01 with respect to the control value (M). Effect of iron on mineralization To investigate whether the inhibiting effect of SNP on mineralization was generated by releasing its iron, the effect of chelating iron was investigated by deferoxamine. Deferoxamine (200 µM) was able to reverse the inhibiting effect of SNP on mineralization (Fig. 5). Based on this result it was investigated whether free iron ions released by ammonium iron (II) sulfate (Fe2+) and/or ammonium iron (III) sulfate (Fe3+) were able to inhibit mineralization. Fig. 5 shows that both agents inhibit mineralization. To assess whether SNP was able to release its iron under our culture conditions, we measured the iron concentration in the cell culture medium. Fourteen days cultured cells were incubated for 24 h in mineralization medium with addition of 5 mM Pi + 100 µM SNP. Free or weakly complexed iron (Fe2+) was determined using Ferrozine and measured to be 30 ± 6 µM (mean ± S.E.M., n = 3). No free Fe2+ could be detected in medium containing SNP without cells. Altogether, these data indicate that the inhibiting effect of mineralization by SNP is generated by inorganic iron derived from SNP. Next, the dose-response effect of SNP, Fe2+ and Fe3+ on mineralization was investigated. Fig. 6 shows that SNP, Fe2+ and Fe3+ already inhibit mineralization more than 60% at a concentration of 10 µM. Maximum inhibition was measured at an iron concentration of 30 µM, which is in line with the amount of iron released from 100 µM SNP.

Chapter 5

102

Figure 5. Effect of iron ions on mineralization. ATDC5 cells were grown 14 days in differentiation medium and subsequently 24 h in mineralization medium with addition of 5 mM Pi (M), or 5 mM Pi + 100 µM sodium nitroprusside (SNP), or 5 mM Pi + 100 µM ammonium iron (II) sulfate (Fe2+) , or 5mM Pi + 100 µM ammonium iron (III) sulfate (Fe3+), or 5 mM Pi + 100 µM SNP + 200 µM deferoxamine (DFO). Cell cultures were fixed in 70% ethanol and stained with 0.5% Alizarin red. Subsequently the alizarin red stain was extracted and measured by a spectrophotometer as described in the Materials and Methods section. Values are mean ± S.E.M (n=3), * P< 0.01 with respect to the control value (M). Of note, ammonium iron sulfate (≥ 100 µM) interfered non specifically with the alizarin red stain, resulting in apparent higher alizarin red stain.

Figure 6. Dose-response curve of SNP, Fe2+ and Fe3+ on mineralization. ATDC5 cells were grown 14 days in differentiation medium and subsequently 24 h in mineralization medium with addition of 5 mM Pi and different concentrations SNP (○), ammonium iron (II) sulfate (Fe2+) (■) and ammonium iron (III) sulfate (Fe3+) (∆). Cell cultures were fixed in 70% ethanol and stained with 0.5% Alizarin red. Subsequently the alizarin red stain was extracted and measured by a spectrophotometer as described in the Materials and Methods section. Results are expressed as mean ± S.E.M.


103

Involvement of radicals in effect of sodium nitroprusside SNP is a redox iron complex that can generate reactive oxygen species including hydroxyl radicals and lipid radicals [19]. This effect can be induced by the iron moiety of breakdown products of SNP, inducing the so called Fenton reaction [19]. To investigate whether SNP inhibited mineralization by generation of reactive oxygen species, the effect of super oxide dismutase (350 U/ml), catalase (6000 U/ml) and mannitol (30 mM) on the inhibiting effect was determined. As shown in Fig. 7 both enzymes were not able to antagonize the effect of SNP on mineralization.

Figure 7. Radical scavengers do not prevent the inhibiting effect of SNP on mineralization. ATDC5 cells were grown 14 days in differentiation medium and subsequently 24 h in mineralization medium with addition of 5 mM Pi (M), or 5 mM Pi + 100 µM SNP (SNP), or 5 mM Pi + 100 µM SNP + 30 mM mannitol (Man), or 5 mM Pi + 100 µM SNP + 6000 U/ml catalase (CAT), or 5 mM Pi + 100 µM SNP + 350 U/ml superoxide dismutase (SOD). Cell cultures were fixed in 70% ethanol and stained with 0.5% Alizarin red. Subsequently the alizarin red stain was extracted and measured by a spectrophotometer as described in the Materials and Methods section. The results represent the means ± S.E.M. of three independent experiments each performed in duplo, * P< 0.01 with respect to the control value (M). Effect of bepridil on the inhibiting effect of sodium nitroprusside SNP has been shown to activate the Na+-Ca2+ exchanger by decomposition to iron in C6 glioma cells. This activating effect was antagonized by the exchange inhibitor bepridil [20;21]. Therefore it was investigated whether bepridil (10 µM) could reverse the inhibiting effect of SNP on mineralization. However, bepridil was not able to antagonize the inhibiting effect generated by SNP (Fig. 8).

Chapter 5

104

Figure 8. Bepridil could not reverse the inhibiting effect of SNP on mineralization. ATDC5 cells were grown 14 days in differentiation medium and subsequently 24 h in mineralization medium with addition of 5 mM Pi (M), or 5 mM Pi + 100 µM SNP (SNP) or 5 mM Pi + 100 µM SNP + 10 µM bepridil (SNP + BEP), or 5 mM Pi + 10 µM bepridil (BEP). Cell cultures were fixed in 70% ethanol and stained with 0.5% Alizarin red. Subsequently the alizarin red stain was extracted and measured by a spectrophotometer as described in the Materials and Methods section. The results represent the means ± S.E.M. of three independent experiments each performed in duplo, * P< 0.01 with respect to the control value (M). Discussion This study demonstrates that although SNP is generally described as an NO donating drug, its inhibiting effect on mineralization is not mediated by NO, but rather by releasing its iron. SNP is a metal nitrosyl complex, which consists of iron, cyanide groups and a nitro moiety (Fig. 1) [15]. Decomposition compounds generated from SNP other than NO, have been reported to be responsible for a number of cellular effects [16;18-20;22-25]. Decomposition of SNP to the toxic compound cyanide has been reported [25-27]. Cyanide represents 44% w/w of the sodium nitroprusside molar mass [26;28]. The toxic effects of cyanide have been attributed to inhibition of cytochrome c oxidase, the terminal enzyme of the respiratory chain, which compromises oxidative phosphorylation leading to cytotoxic hypoxia [26]. This study shows, however, that SNP does not inhibit mineralization by decomposition to cyanide, as 100 µM potassium cyanide was not able to inhibit mineralization. Another degradation product of SNP is hexacyanoferrate. Ferrocyanide has been reported to be a moderately active antagonist of the N-methyl-D-aspartate receptor function in cerebellar cells [23;29], and ferricyanide has been shown to activate Na+-Ca2+ exchanger in C6 glioma cells [20]. Ferrocyanide and ferricyanide, in contrast to ammonium iron sulfate however, did not inhibit mineralization in our


105

cell culture system. This may be explained by their kinetic stability, preventing the iron moiety to react with other compounds. Ammonium iron sulfate, which decomposes in water to free iron ions, was able to inhibit mineralization at a concentration of 30 µM in our cell culture system. This concentration of iron was also measured to be released from 100 µM SNP under our culture conditions. SNP has been shown to produce oxygen radicals, which can react with hydrogen to hydrogen peroxide [16]. Fe2+ promotes oxidative stress through reactive oxygen species formation [30]. It is therefore possible that Fe2+ inhibits mineralization by generating hydroxyl radicals by reduction of hydrogen peroxide by the Fenton reaction [31]. Hydroxyl radicals are the most reactive free radicals and can induce lipid peroxidation and cytotoxity [19;32]. Therefore it was investigated whether the inhibiting effect on mineralization, caused by the iron moiety of SNP, was mediated by hydrogen radicals. The possible involvement of hydroxyl radicals was examined by testing the ability of a number of radical scavengers to antagonize the effect of SNP [33]. Catalase, SOD and mannitol were not able to antagonize SNP, indicating that the inhibiting effect was not mediated by induction of hydroxyl radicals. Decomposition of SNP to iron has been reported to stimulate calcium influx by activation of Na+-Ca2+ exchanger preventing chemical hypoxia induced cell death. This effect was antagonized by the Na+-Ca2+ exchanger inhibitor, bepridil [20]. However, long term exposure to bepridil has been shown to significantly inhibit mineralization in primary osteoblasts [34]. In this study, it was found that bepridil, when added for 24 h to the mineralization medium, could neither inhibit mineralization nor antagonize the inhibiting effect on mineralization generated by SNP, excluding a role the Na+-Ca2+ exchanger in the effect of SNP on mineralization. Taken altogether, the results obtained in this study suggest that the iron ions generated by SNP inhibit mineralization via another unknown mechanism. Low concentrations of Fe3+ have been shown to inhibit cardiovascular calcification induced by bioprosthetic heart valves [35-37]. Calcification accounts for 60% of implanted cardiac tissue failure [38;39]. Co-implantation with FeCl3 loaded chitosan beads has been shown to be an effective treatment [35;40-42]. In these studies it was assumed that ferric ions retard the calcification process by inhibiting proper formation of hydroxyapatite [35;37]. Ferric compounds have also been shown to be effective dietary phosphate binders [36]. The majority of the studies concerning the influence of ferric compounds on phosphate binding have utilized animal studies and focused mainly on the restriction of intestinal phosphate absorption to control hyperphosphatemia in patients with end stage renal disease. The estimated binding capacity of ferric compounds ranged from 88 to 181 mg phosphate per gram elemental iron [36;43].

Chapter 5

106

This indicates that the iron concentration has to be 10 times higher than the phosphate concentration for its binding capacity. Under our cell culture conditions, however, the iron concentration is at least 60 times lower than the phosphate concentration. This indicates that it is unlikely that in our cell culture system iron works as a phosphate binder. To our knowledge no effects of iron ions on mineralization in vitro have been described. Other metal ions have been reported to inhibit mineralization in vitro. For example, aluminum has been reported to inhibit mineralization in vitro, as well as manganese and fluoride at concentrations of 10 µM or greater. Depending on the ionic status, these metals were expected to be strong inhibitors of calcium phosphate nucleation and crystal growth [44-48]. Zinc has also been shown to inhibit matrix vesicles-mediated mineralization, at concentrations ranging from 5-100 µM [49;50]. It was speculated that the inhibitory effect of zinc was either through blocking the intravesicular Ca2+-Pi-phospholipid complexes or through inhibiting the function of annexin channels, which are responsible for calcium transport [49;50]. Other investigators have suggested that once mineral nucleation has occurred zinc ions may become incorporated into the hydroxyapatite crystal, and interfere with crystal growth [51]. The inhibitory effect of zinc at a concentration of 100 µM on mineralization was also found in our cell culture system (data not shown). We therefore speculate that iron ions inhibit mineral formation in our cell culture system by a similar mechanism as proposed for zinc e.g. by interfering with initial entry of calcium ions through annexin channels or by preventing hydroxyapatite crystal nucleation. In summary, this study shows that decomposition of SNP to iron ions inhibit mineralization. SNP is used therapeutically in heart transplantations for potent vasodilating effects [52-54]). Based on the results of this study, it seems likely that this treatment also induces inhibition of vascular calcification by the release of the iron moiety of SNP. Reference List 1. Heinegard D, Oldberg A (1989) Structure and biology of cartilage and bone matrix

noncollagenous macromolecules. FASEB J. 3:2042-2051 2. Kronenberg HM (2003) Developmental regulation of the growth plate. Nature 423:332-336 3. Karsenty G, Wagner EF (2002) Reaching a genetic and molecular understanding of skeletal

development. Dev.Cell 2:389-406 4. Giachelli CM (2005) Inducers and inhibitors of biomineralization: lessons from

pathological calcification. Orthod.Craniofac.Res. 8:229-231 5. Speer MY, Giachelli CM (2004) Regulation of cardiovascular calcification.

Cardiovasc.Pathol. 13:63-70 6. Anderson HC (1995) Molecular biology of matrix vesicles. Clin.Orthop.Relat Res.266-280


107

7. Hsu HH, Camacho NP, Sun F, Tawfik O, Aono H (2000) Isolation of calcifiable vesicles from aortas of rabbits fed with high cholesterol diets. Atherosclerosis 153:337-348

8. Hsu HH, Tawfik O, Sun F (2002) Mechanisms of calcification by vesicles isolated from atherosclerotic rabbit aortas. Biochim.Biophys.Acta 1563:18-22

9. Hsu HH, Camacho NC, Tawfik O, Sun F (2002) Induction of calcification in rabbit aortas by high cholesterol diets: roles of calcifiable vesicles in dystrophic calcification. Atherosclerosis 161:85-94


11. Bobryshev YV (2005) Transdifferentiation of smooth muscle cells into chondrocytes in atherosclerotic arteries in situ: implications for diffuse intimal calcification. J.Pathol. 205:641-650

12. Bostrom K, Watson KE, Horn S, Wortham C, Herman IM, Demer LL (1993) Bone morphogenetic protein expression in human atherosclerotic lesions. J.Clin.Invest 91:1800-1809

13. Huitema LF, Vaandrager AB, van Weeren PR, Barneveld A, Helms JB, van de Lest CH (2006) The nitric oxide donor sodium nitroprusside inhibits mineralization in ATDC5 cells. Calcif.Tissue Int. 78:171-177

14. Butler AR, Megson IL (2002) Non-heme iron nitrosyls in biology. Chem.Rev. 102:1155-1166

15. Roncaroli F, van ER, Olabe JA (2005) Release of NO from reduced nitroprusside ion. Iron-dinitrosyl formation and NO-disproportionation reactions. Inorg.Chem. 44:2781-2790

16. Terwel D, Nieland LJ, Schutte B, Reutelingsperger CP, Ramaekers FC, Steinbusch HW (2000) S-nitroso-N-acetylpenicillamine and nitroprusside induce apoptosis in a neuronal cell line by the production of different reactive molecules. Eur.J.Pharmacol. 400:19-33

17. Bates JN, Baker MT, Guerra R, Jr., Harrison DG (1991) Nitric oxide generation from nitroprusside by vascular tissue. Evidence that reduction of the nitroprusside anion and cyanide loss are required. Biochem.Pharmacol. 42 Suppl:S157-S165

18. Rao DN, Elguindi S, O'Brien PJ (1991) Reductive metabolism of nitroprusside in rat hepatocytes and human erythrocytes. Arch.Biochem.Biophys. 286:30-37

19. Rauhala P, Khaldi A, Mohanakumar KP, Chiueh CC (1998) Apparent role of hydroxyl radicals in oxidative brain injury induced by sodium nitroprusside. Free Radic.Biol.Med. 24:1065-1073

20. Amoroso S, Tortiglione A, Secondo A, Catalano A, Montagnani S, Di RG, Annunziato L (2000) Sodium nitroprusside prevents chemical hypoxia-induced cell death through iron ions stimulating the activity of the Na+-Ca2+ exchanger in C6 glioma cells. J.Neurochem. 74:1505-1513

21. Stys PK, Waxman SG, Ransom BR (1991) Na(+)-Ca2+ exchanger mediates Ca2+ influx during anoxia in mammalian central nervous system white matter. Ann.Neurol. 30:375-380

22. Jaworski K, Kinard F, Goldstein D, Holvoet P, Trouet A, Schneider YJ, Remacle C (2001) S-nitrosothiols do not induce oxidative stress, contrary to other nitric oxide donors, in cultures of vascular endothelial or smooth muscle cells. Eur.J.Pharmacol. 425:11-19

23. Neijt H, Koller M, Urwyler S (2001) Inorganic iron complexes derived from the nitric oxide donor nitroprusside: competitive N-methyl-D-aspartate receptor antagonists with nanomolar affinity. Biochem.Pharmacol. 61:343-349

24. Serra PA, Rocchitta G, Esposito G, Delogu MR, Migheli R, Miele E, Desole MS, Miele M (2001) A study on the role of nitric oxide and iron in 3-morpholino-sydnonimine-induced increases in dopamine release in the striatum of freely moving rats. Br.J.Pharmacol. 134:275-282

Chapter 5

108

25. Spiegel HE, Kucera V (1977) Some aspects of sodium nitroprusside reaction with human erythrocytes. Clin.Chem. 23:2329-2331

26. Jensen MS, Nyborg NC, Thomsen ES (2000) Various nitric oxide donors protect chick embryonic neurons from cyanide-induced apoptosis. Toxicol.Sci. 58:127-134

27. Michenfelder JD (1977) Cyanide release from sodium nitroprusside in the dog. Anesthesiology 46:196-201

28. Schulz V (1984) Clinical pharmacokinetics of nitroprusside, cyanide, thiosulphate and thiocyanate. Clin.Pharmacokinet. 9:239-251

29. East SJ, Batchelor AM, Garthwaite J (1991) Selective blockade of N-methyl-D-aspartate receptor function by the nitric oxide donor, nitroprusside. Eur.J.Pharmacol. 209:119-121

30. Jellinger KA (1999) The role of iron in neurodegeneration: prospects for pharmacotherapy of Parkinson's disease. Drugs Aging 14:115-140

31. McCord JM, Day ED, Jr. (1978) Superoxide-dependent production of hydroxyl radical catalyzed by iron-EDTA complex. FEBS Lett. 86:139-142

32. Halliwell B, Gutteridge JM (1992) Biologically relevant metal ion-dependent hydroxyl radical generation. An update. FEBS Lett. 307:108-112

33. Starke PE, Farber JL (1985) Ferric iron and superoxide ions are required for the killing of cultured hepatocytes by hydrogen peroxide. Evidence for the participation of hydroxyl radicals formed by an iron-catalyzed Haber-Weiss reaction. J.Biol.Chem. 260:10099-10104

34. Stains JP, Gay CV (2001) Inhibition of Na+/Ca2+ exchange with KB-R7943 or bepridil diminished mineral deposition by osteoblasts. J.Bone Miner.Res. 16:1434-1443

35. Chandy T, Mohanty M, John A, Rao SB, Sivakumar R, Sharma CP, Valiathan MS (1996) Structural studies on bovine bioprosthetic tissues and their in vivo calcification: prevention via drug delivery. Biomaterials 17:577-585

36. Hsu CH, Patel SR, Young EW (1999) New phosphate binding agents: ferric compounds. J.Am.Soc.Nephrol. 10:1274-1280

37. Vasudev SC, Chandy T, Umasankar MM, Sharma CP (2001) Inhibition of bioprosthesis calcification due to synergistic effect of Fe/Mg ions to polyethylene glycol grafted bovine pericardium. J.Biomater.Appl. 16:93-107

38. Schoen FJ, Levy RJ (1984) Bioprosthetic heart valve failure: pathology and pathogenesis. Cardiol.Clin. 2:717-739

39. Silver M, Hudson RE, Trimble AS (1975) Morphologic observations on heart valve prostheses made of fascia lata. J.Thorac.Cardiovasc.Surg. 70:360-366

40. Chandy T, Sharma CP (1996) Effect of liposome-albumin coatings on ferric ion retention and release from chitosan beads. Biomaterials 17:61-66

41. Pathak YV, Boyd J, Levy RJ, Schoen FJ (1990) Prevention of calcification of glutaraldehyde pretreated bovine pericardium through controlled release polymeric implants: studies of Fe3+, Al3+, protamine sulphate and levamisole. Biomaterials 11:718-723

42. Tan WM, Loke WK, Tan BL, Wee A, Khor E, Goh KS (1993) Trivalent metal ions in the prevention of calcification in glutaraldehyde treated biological tissues. Is there a chemical correlation? Biomaterials 14:1003-1007

43. Zhu H, Webb M, Buckley J, Roberts NB (2002) Different Mg to Fe ratios in the mixed metal MgFe hydroxy-carbonate compounds and the effect on phosphate binding compared with established phosphate binders. J.Pharm.Sci. 91:53-66

44. Bellows CG, Aubin JE, Heersche JN (1993) Differential effects of fluoride during initiation and progression of mineralization of osteoid nodules formed in vitro. J.Bone Miner.Res. 8:1357-1363


109

45. Bellows CG, Aubin JE, Heersche JN (1995) Aluminum inhibits both initiation and progression of mineralization of osteoid nodules formed in differentiating rat calvaria cell cultures. J.Bone Miner.Res. 10:2011-2016

46. Litchfield TM, Ishikawa Y, Wu LN, Wuthier RE, Sauer GR (1998) Effect of metal ions on calcifying growth plate cartilage chondrocytes. Calcif.Tissue Int. 62:341-349

47. Sprague SM, Krieger NS, Bushinsky DA (1993) Aluminum inhibits bone nodule formation and calcification in vitro. Am.J.Physiol 264:F882-F890

48. Talwar HS, Reddi AH, Menczel J, Thomas WC, Jr., Meyer JL (1986) Influence of aluminum on mineralization during matrix-induced bone development. Kidney Int. 29:1038-1042

49. Kirsch T, Harrison G, Worch KP, Golub EE (2000) Regulatory roles of zinc in matrix vesicle-mediated mineralization of growth plate cartilage. J.Bone Miner.Res. 15:261-270

50. Sauer GR, Wu LN, Iijima M, Wuthier RE (1997) The influence of trace elements on calcium phosphate formation by matrix vesicles. J.Inorg.Biochem. 65:57-65

51. Mayer I, Apfelbaum F, Featherstone JD (1994) Zinc ions in synthetic carbonated hydroxyapatites. Arch.Oral Biol. 39:87-90

52. Kaplan JA, Finlayson DC, Woodward S (1980) Vasodilator therapy after cardiac surgery: a review of the efficacy and toxicity of nitroglycerin and nitroprusside. Can.Anaesth.Soc.J. 27:254-259

53. Ryan LM, Cheung HS, LeGeros RZ, Kurup IV, Toth J, Westfall PR, McCarthy GM (1999) Cellular responses to whitlockite. Calcif.Tissue Int. 65:374-377

54. Costard-Jackle A, Fowler MB (1992) Influence of preoperative pulmonary artery pressure on mortality after heart transplantation: testing of potential reversibility of pulmonary hypertension with nitroprusside is useful in defining a high risk group. J.Am.Coll.Cardiol. 19:48-54

Chapter 6

Summarizing discussion

Chapter 6

112

1. Introduction While ectopic calcifications have long been considered to result from a mere physiochemical precipitation of calcium and phosphate (Pi), recent studies have provided evidence that they may actually share many molecular and cellular features with bone and cartilage mineralization [1]. An increasing amount of findings suggest that vascular smooth muscle cells promote calcification by mechanisms very similar to those occurring during growth plate mineralization [1]. Vascular calcification is a manifestation of atherosclerosis and atherosclerosis is the major cause of mortality in the western hemisphere. For now it is unclear why atherosclerotic lesions have the tendency to calcify [2;3]. In addition, calcification of articular cartilage contributes to significant morbidity [4]. By improving our knowledge of the complex process of mineralization and in particular about the beginning stages of mineralization, we might be better equipped to design rational treatments for aging diseases like atherosclerosis and osteoarthritis. Many investigations have contributed greatly to a basic understanding of the mechanisms regulating cell-mediated mineralization, either with a focus on the formation of skeletal tissue, or with a focus on pathological soft tissue mineralization. The use of cell culture systems made it possible to investigate mineralization at a cellular level. There is still much debate regarding the biochemical mechanisms that initiate mineralization subsequent to the increase in calcium and/or Pi. As described in chapter 1, one theory proposes that mineralization is initiated within matrix vesicles (MVs) [5]. A second not mutually exclusive theory, suggests that Pi induces apoptosis, and that apoptotic bodies nucleate calcium and Pi [6;7]. A third theory proposes that mineralization is mediated by soluble phosphoprotein nucleators, which associate with the extracellular matrix [8-10]. Regardless of the way mineralization is initiated, vertebrates may also actively inhibit mineralization by secreting specific proteins. Therefore removal or inactivation of an inhibitor may also promote mineralization [11]. In view of on these different hypotheses, the aim of this study was to gain more insight in the processes that take place during the initiation of cell-mediated mineralization (chapters 2 and 3) and the effectors involved in these processes (chapters 4 and 5). The mouse embryonal carcinoma-derived ATDC5 cell line was used in our studies, since it is known to be a suitable model to investigate cell-mediated mineralization [12;13].


113

2. How is cell-mediated mineralization initiated in ATDC5 cells? As described in chapter 2, ATDC5 cells form large mineralizing structures (LMS) in 2 hours in medium containing additional Pi in the absence of serum. Serum is known to contain the mineralization inhibitor fetuin, which may explain the relatively short time period required for mineral crystal formation [14-16]. Excluding serum from the medium also shortened the cell culture period, since LMS formation is independent of the presence of an extracellular matrix. The hydroxyapatite crystals (or LMS) in the serum-free cell culture system were able to expand rapidly to sub millimeter dimensions in the medium, which to our knowledge has never been reported before. LMS were shown to contain cells, which were observed to be stretched apart and integrated in the crystal. If the mineral crystals had developed onto extracellular matrix proteins, we would not expect cells to be stretched apart. Since non-collagenous phosphoproteins are generally extracellular matrix proteins, the formation of LMS does not fit in the theory of nucleating non-collagenous phosphoproteins (third theory) [8-10]. By that same reasoning we suggest that these crystals also did not develop in extracellular vesicles, like MVs (first theory) or apoptotic bodies (second theory). For comparison, mineralizing structures observed in staurosporin-treated cells, which may be nucleated in secreted apoptotic bodies [17-19], do not resemble LMS and do not contain disordered nuclei. We speculate that the most likely explanation for the formation of LMS is that various crystal nucleation cores are formed inside the cell in different subcellular locations and subsequently grow into different directions, stretching the cell apart. The presence of multiple mineralization cores inside a cell was indeed observed under circumstances that the mineralization process was stalled by the inclusion of serum in the medium. Based on these observations, we suggest an alternative theory of the initiation of mineralization, which is that the first crystal formation may start intracellularly as proposed by Stanford et al. (1995) [20]. Of note, we could only clearly observe intracellular crystals in apoptotic or leaky cells (e.g. Fig. 7 chapter 2 and Fig. 2 appendix A), which suggest that Pi-induced mineralization results in cell death. Since mineral crystals may be toxic to the cell, it may also be possible that the cells subsequently secrete the intracellularly formed crystals, as a protective mechanism to remove the crystals. In view of the suggestion that mineralization may start intracellularly, it is intriguing that annexin V was translocated from the cytosol to the membrane after addition of Pi. This translocation was partly inhibited by the Pi transporter inhibitor phosphonoformic acid in parallel with inhibition of LMS formation. Translocation of annexin V to the membrane has been shown to be calcium dependent and reported to result in the formation of calcium transport channels in a PS rich membrane [21-23]. This suggests that Pi may induce calcium transport into the cells, and subsequently crystals may develop intracellularly.

Chapter 6

114

Therefore, in chapter 2, appendix B cytosolic free calcium was measured after addition of Pi. However, the cytosolic free calcium concentration was not significantly increased in the ATDC5 cells after addition of Pi compared to the control cells (without addition of Pi). Explanations for this result may be that Pi induced Annexin V translocation is mediated via another pathway than cytosolic calcium influx or that Pi raises calcium concentrations only locally close to the membrane. The cytoplasmic free calcium is normally maintained at about 0.1 µM, which is very low relative to the extracellular fluid ([Ca2+] =1.2 mM) [24]. Therefore, it seems likely that for the intracellular induction of mineralization calcium and Pi are compartmentalized in an intracellular organelle. Such an organelle must have the capacity to increase its calcium and Pi concentration. Furthermore, distortion of this particular organelle is likely to result in apoptotic cells, because we observed calcium containing crystals predominantly in apoptotic cells (chapter 2, appendix A). Since intracellular concentrations of calcium and Pi are regulated by mitochondria and mitochondrial dysfunction is known to trigger apoptosis, mitochondria may be candidates for the formation of the first crystal nuclei in ATDC5 cells [25;26]. Pi influx into the mitochondria is necessary during oxidative phosphorylation for conversion of adenosine diphosphate (ADP) to adenosine triphosphate (ATP) [27]. Calcium influx is required to activate certain hydrogenases coupled to the Krebs cycle [25]. In addition, the import of calcium into the mitochondria may also be important for removing calcium from the cytosol [26]. When the calcium concentration in the mitochondria becomes too high it may induce the irreversible opening of the permeability transition pore (PTP). This will then result in calcium release from mitochondria, collapse of the membrane potential and uncoupled ATP synthesis [26]. It has been reported that addition of Pi (> 1 mM) to cultured chondrocytes causes a decrease in mitochondrial membrane potential, which indicates that Pi signals to the mitochondria [6;28]. In addition, it has been shown that elevated cytosolic Pi can induce mitochondrial membrane permeability transition [29]. In this altered state, proteins on the inner and outer mitochondrial membranes dissociate with a loss of selective permeability, resulting to the death pathway [30;31]. However, the free calcium concentration in mitochondria is normally approximately 40 µM [32], which is possibly too low to complex with Pi, unless other unknown nucleating factors, such as increase in pH, exclusion of mineralization inhibitors, or specific nucleating proteins are present in the mitochondria. Alternatively, the first mineral crystals may form in the endoplasmatic reticulum (ER) in ATDC5 cells. The ER is the most important calcium storage compartment in most cells and acts as an intracellular calcium buffer, which is important for the removal of excess cytosolic calcium that may be accumulated during physiological


115

stimulation [24;33]. Calcium is sequestered into the ER by ATP consuming pumps known as the sarco-endoplasmic reticulum calcium ATPases (SERCAs) [33]. The calcium concentration varies in the ER between 0.1 – 1 mM [24;34]. In addition, emerging evidence suggests that the ER also regulates apoptosis, since calcium release from the ER, has been implicated as a key regulator event in many apoptotic models [35]. Depending on the mode of calcium release it may either directly activate death effectors or influence the sensitivity of mitochondria to apoptotic transitions [35]. Based on these properties, the ER is also a candidate for the first mineral crystal formation. However, little is known about the Pi concentration in the ER and eventually Pi transport into the ER lumen when cytosolic Pi is increased. It may also be possible that ATDC5 cells precipitate calcium and Pi by taking up tissue fluid by endocytosis or pinocytosis (cellular drinking). Pinocytosis involves the ingestion of tissue fluid and solutes via small vesicles (≤ 150 nm) [36]. These vesicles then contain tissue fluid with 1.2 mM calcium and an elevated Pi concentration. Subsequently, molecules in these vesicles may be delivered to early endosomes were mineralization may start. Alternatively, endosomes may fuse with lysosomes and it may be possible that calcium and Pi ions are sequestered in lysosomes, which may subsequently form residual bodies [36]. Calcium Pi precipitation in these subcellular compartments may be normally inhibited due to the acidic pH in the lumen (lysosomal pH ~5). This acidic pH is generated by an H+ ATPase in the lysosomal membrane that pumps H+ into the lumen [36]. When for any reason (Pi-induced) these pumps are not functioning anymore, the subsequent increased pH may result in calcium Pi precipitation in the lysosomes/residual bodies. 3. Effectors involved in the mineralization process 3.1 Factors that affect the formation of calcium phosphate crystals The results obtained in the study described in chapter 3 show that soluble factors released by ATDC5 cells can affect the formation of calcium Pi crystals. In this study, subconfluent cultured ATDC5 cells were incubated for 1 hour in medium without serum. Subsequently, medium was harvested and incubated for 24 hours in the presence of additional Pi. This resulted in the formation of flat mineralizing structures (FMS), consisting of complex irregularly shaped flat crystals, which occasionally contained fiber-like structures. Without pre-incubation of medium with cells, only small punctuate calcium Pi crystals formed. The formation of FMS was shown to be caused by soluble factures constitutively released by ATDC5 cells. FMS formation is different from LMS formation described in chapter 2, since FMS do not contain cells. Furthermore, the incubation period for mineral crystal

Chapter 6

116

formation was 24 hours for FMS versus 2 hours for LMS and LMS (sub millimetre size range) were bigger than FMS (approximately 40 µm in size).In the 2 hour time period required for LMS formation calcium Pi precipitation did not occur in medium without cells, in contrast to the 24 hour time period required for FMS formation. This may suggest that LMS contain cell specific crystal nucleating factors. The secreted factors involved in FMS formation may possibly play a role in the growth phase of LMS formation once in the initial phase the crystal nucleus has been formed. 3.2 Effect of the nitric oxide donor drug sodium nitroprusside on mineralization In our search for agents that can affect mineralization we found that the NO donor sodium nitroprusside (SNP) inhibits mineralization by more than 60% at a concentration of 10 µM. This is described in chapter 4. The relevance of this study is that SNP is used therapeutically in heart transplantations for its potent vasodilating effects [37-39]. Based on the results of this study, it seems likely that this treatment also induces inhibition of vascular calcification. However, the inhibition of mineralization by SNP was not affected by inhibitors of guanylyl cyclase nor mimicked by a cGMP analog, as observed in the case of vasodilating effect of SNP [40]. Furthermore, sodium nitroprusside did not inhibit Pi uptake nor inhibited apoptosis in the ATDC5 cells. Therefore, a follow up study was performed in chapter 5, which led to the conclusion that, although SNP is generally described as an NO donating drug in biological and medical literature, its inhibiting effect on mineralization is not mediated by NO, but rather by releasing iron. To our knowledge no effects of iron ions on mineralization in vitro have been described. Other metal ions have been reported to inhibit mineralization in vitro. For example, aluminum has been reported to inhibit mineralization in vitro, as well as manganese, fluoride and zinc at concentrations of 10 µM or greater [41-46]. An inhibitory effect of zinc at a concentration of 100 µM on mineralization was also found in our cell culture system (data not shown). The mechanism by which iron inhibits mineral formation is as yet not known. Possibly iron may become incorporated into the hydroxyapatite crystal, and interfere with crystal growth [46]. Alternatively, iron might block the intravesicular Ca2+-Pi-phospholipid complexes or through inhibiting the function of annexin channels [47;48]. 4. Future directions Although we observed intracellular mineral crystals in our ATDC5 model system (chapter 2 and appendix A to chapter 2) we were unable to define in which subcellular compartment the initial crystals develop. Therefore, an interesting future direction is to investigate whether and in which subcellular compartment the


117

first mineral crystal forms. This can be performed by live cell calcium imaging of organelles like the ER, mitochondria and lysosomes [32;34]. It is also interesting to perform live cell imaging studies with a pH indicator to investigate whether a local intracellular increase in pH is occurring during the induction of mineralization. Electron microscopy studies with antibodies for the putative mineralizing organelles at earlier stages of mineralization may also clarify in which subcellular compartment mineralization is initiated. Alternatively, it is worth trying to isolate the calcein labeled mineral crystals by a fluorescent-activated cell sorter (FACS) and subsequently analyze the proteome. Using this approach an interesting nucleating protein might also be identified. To investigate which soluble factors released by ATDC5 cells affect the formation of calcium Pi crystals, one can fractionate the soluble proteins released by ATDC5 cells by column chromatography. Fractionated proteins can then be tested for the effect on calcium Pi formation in the model described in chapter 3. Alternatively, it is interesting to investigate whether the soluble factors, contain other factors than proteins, like for example phospholipids. Based on the results of the studies in chapters 4 and 5, it seems that vasodilating therapeutic treatment with SNP during heart transplantations also induces inhibition of vascular calcification. Calcification accounts for 60% of implanted cardiac tissue failure [49;50]. An interesting future direction is to investigate the therapeutic use of SNP to inhibit calcification in cardiac tissue transplantation. Reference List 1. Magne D, Julien M, Vinatier C, Merhi-Soussi F, Weiss P, Guicheux J (2005) Cartilage

formation in growth plate and arteries: from physiology to pathology. Bioessays 27:708-716 2. Abedin M, Tintut Y, Demer LL (2004) Vascular calcification: mechanisms and clinical

ramifications. Arterioscler.Thromb.Vasc.Biol. 24:1161-1170 3. Schinke T, Amendt C, Trindl A, Poschke O, Muller-Esterl W, Jahnen-Dechent W (1996)

The serum protein alpha2-HS glycoprotein/fetuin inhibits apatite formation in vitro and in mineralizing calvaria cells. A possible role in mineralization and calcium homeostasis. J.Biol.Chem. 271:20789-20796

4. Rutsch F, Terkeltaub R (2005) Deficiencies of physiologic calcification inhibitors and low-grade inflammation in arterial calcification: lessons for cartilage calcification. Joint Bone Spine 72:110-118

5. Anderson HC (1995) Molecular biology of matrix vesicles. Clin.Orthop.Relat Res.266-280 6. Shapiro IM, Adams CS, Freeman T, Srinivas V (2005) Fate of the hypertrophic

chondrocyte: microenvironmental perspectives on apoptosis and survival in the epiphyseal growth plate. Birth Defects Res.C.Embryo.Today 75:330-339

7. Gibson G (1998) Active role of chondrocyte apoptosis in endochondral ossification. Microsc.Res.Tech. 43:191-204

Chapter 6

118

8. Glimcher MJ (1989) Mechanism of calcification: role of collagen fibrils and collagen-phosphoprotein complexes in vitro and in vivo. Anat.Rec. 224:139-153



11. Schinke T, Karsenty G (2000) Vascular calcification--a passive process in need of inhibitors. Nephrol.Dial.Transplant. 15:1272-1274


13. Shukunami C, Shigeno C, Atsumi T, Ishizeki K, Suzuki F, Hiraki Y (1996) Chondrogenic differentiation of clonal mouse embryonic cell line ATDC5 in vitro: differentiation-dependent gene expression of parathyroid hormone (PTH)/PTH-related peptide receptor. J.Cell Biol. 133:457-468



16. Schafer C, Heiss A, Schwarz A, Westenfeld R, Ketteler M, Floege J, Muller-Esterl W, Schinke T, Jahnen-Dechent W (2003) The serum protein alpha 2-Heremans-Schmid glycoprotein/fetuin-A is a systemically acting inhibitor of ectopic calcification. J.Clin.Invest 112:357-366




20. Stanford CM, Jacobson PA, Eanes ED, Lembke LA, Midura RJ (1995) Rapidly forming apatitic mineral in an osteoblastic cell line (UMR 106-01 BSP). J Biol.Chem. 270:9420-9428


22. Barwise JL, Walker JH (1996) Annexins II, IV, V and VI relocate in response to rises in intracellular calcium in human foreskin fibroblasts. J.Cell Sci. 109 ( Pt 1):247-255

23. Mohiti J, Caswell AM, Walker JH (1995) Calcium-induced relocation of annexins IV and V in the human osteosarcoma cell line MG-63. Mol.Membr.Biol. 12:321-329

24. Hajnoczky G, Davies E, Madesh M (2003) Calcium signaling and apoptosis. Biochem.Biophys.Res Commun. 304:445-454

25. Gunter TE, Buntinas L, Sparagna G, Eliseev R, Gunter K (2000) Mitochondrial calcium transport: mechanisms and functions. Cell Calcium 28:285-296


119

26. Smaili SS, Hsu YT, Youle RJ, Russell JT (2000) Mitochondria in Ca2+ signaling and apoptosis. J Bioenerg.Biomembr. 32:35-46

27. Boyer PD, Stokes BO, Wolcott RG, Degani C (1975) Coupling of "high-energy" phosphate bonds to energy transductions. Fed.Proc. 34:1711-1717

28. Rajpurohit R, Mansfield K, Ohyama K, Ewert D, Shapiro IM (1999) Chondrocyte death is linked to development of a mitochondrial membrane permeability transition in the growth plate. J Cell Physiol 179:287-296

29. Bravo C, Chavez E, Rodriguez JS, Moreno-Sanchez R (1997) The mitochondrial membrane permeability transition induced by inorganic phosphate or inorganic arsenate. A comparative study. Comp Biochem.Physiol B Biochem.Mol.Biol. 117:93-99

30. Susin SA, Zamzami N, Castedo M, Hirsch T, Marchetti P, Macho A, Daugas E, Geuskens M, Kroemer G (1996) Bcl-2 inhibits the mitochondrial release of an apoptogenic protease. J Exp.Med. 184:1331-1341

31. Zamzami N, Marchetti P, Castedo M, Zanin C, Vayssiere JL, Petit PX, Kroemer G (1995) Reduction in mitochondrial potential constitutes an early irreversible step of programmed lymphocyte death in vivo. J Exp.Med. 181:1661-1672

32. Szabadkai G, Simoni AM, Rizzuto R (2003) Mitochondrial Ca2+ uptake requires sustained Ca2+ release from the endoplasmic reticulum. J Biol.Chem. 278:15153-15161

33. Roderick HL, Bootman MD (2005) Redoxing calcium from the ER. Cell 120:4-5 34. Burdakov D, Petersen OH, Verkhratsky A (2005) Intraluminal calcium as a primary

regulator of endoplasmic reticulum function. Cell Calcium 38:303-310 35. Breckenridge DG, Germain M, Mathai JP, Nguyen M, Shore GC (2003) Regulation of

apoptosis by endoplasmic reticulum pathways. Oncogene 22:8608-8618 36. Alberts B, Bray D, Lewis J, Raff M, Roberts K, Watson JD (1994) Vesicular traffic in the

Secretory and Endocytic Pathways. Molecular biology of the cell. pp 599-652 37. Kaplan JA, Finlayson DC, Woodward S (1980) Vasodilator therapy after cardiac surgery: a

review of the efficacy and toxicity of nitroglycerin and nitroprusside. Can.Anaesth.Soc.J. 27:254-259

38. Ryan LM, Cheung HS, LeGeros RZ, Kurup IV, Toth J, Westfall PR, McCarthy GM (1999) Cellular responses to whitlockite. Calcif.Tissue Int. 65:374-377

39. Costard-Jackle A, Fowler MB (1992) Influence of preoperative pulmonary artery pressure on mortality after heart transplantation: testing of potential reversibility of pulmonary hypertension with nitroprusside is useful in defining a high risk group. J.Am.Coll.Cardiol. 19:48-54

40. Lovren F, Triggle C (2000) Nitric oxide and sodium nitroprusside-induced relaxation of the human umbilical artery. Br.J Pharmacol. 131:521-529

41. Bellows CG, Aubin JE, Heersche JN (1993) Differential effects of fluoride during initiation and progression of mineralization of osteoid nodules formed in vitro. J.Bone Miner.Res. 8:1357-1363

42. Bellows CG, Aubin JE, Heersche JN (1995) Aluminum inhibits both initiation and progression of mineralization of osteoid nodules formed in differentiating rat calvaria cell cultures. J.Bone Miner.Res. 10:2011-2016

43. Litchfield TM, Ishikawa Y, Wu LN, Wuthier RE, Sauer GR (1998) Effect of metal ions on calcifying growth plate cartilage chondrocytes. Calcif.Tissue Int. 62:341-349

44. Sprague SM, Krieger NS, Bushinsky DA (1993) Aluminum inhibits bone nodule formation and calcification in vitro. Am.J.Physiol 264:F882-F890

45. Talwar HS, Reddi AH, Menczel J, Thomas WC, Jr., Meyer JL (1986) Influence of aluminum on mineralization during matrix-induced bone development. Kidney Int. 29:1038-1042

Chapter 6

120

46. Mayer I, Apfelbaum F, Featherstone JD (1994) Zinc ions in synthetic carbonated hydroxyapatites. Arch.Oral Biol. 39:87-90

47. Kirsch T, Harrison G, Worch KP, Golub EE (2000) Regulatory roles of zinc in matrix vesicle-mediated mineralization of growth plate cartilage. J.Bone Miner.Res. 15:261-270

48. Sauer GR, Wu LN, Iijima M, Wuthier RE (1997) The influence of trace elements on calcium phosphate formation by matrix vesicles. J.Inorg.Biochem. 65:57-65

49. Schoen FJ, Levy RJ (1984) Bioprosthetic heart valve failure: pathology and pathogenesis. Cardiol.Clin. 2:717-739

50. Silver M, Hudson RE, Trimble AS (1975) Morphologic observations on heart valve prostheses made of fascia lata. J.Thorac.Cardiovasc.Surg. 70:360-366

Samenvatting in het Nederlands


122

Inleiding Mineralisatie in een organisme is het afzetten van kalkzouten (die voornamelijk bestaan uit calcium en fosfaat) in dood of levend weefsel. Mineralisatie of verkalking is essentiëel voor de ontwikkeling van botten en wordt gemediëerd door specifieke cellen in het bot. Botten zijn cruciaal voor het menselijke lichaam en hebben een skelet- (bescherming interne organen, steun en biomechanische beweging), een beenmerg- (hematopoëtisch) en een metabole functie (reservoir mineralen). Normaal gesproken mineraliseren andere organen dan bot niet, maar onder bepaalde pathologische condities mineraliseren organen als gewrichtskraakbeen en bloedvaten toch. Mineralisatie van gewrichtskraakbeen (artrose) leidt tot afbraak van gewrichtskraakbeen, ontsteking (artritis), biomechanische stijfheid en ernstige pijn. Daarentegen leidt mineralisatie van de bloedvaten (aderverkalking) tot vernauwing van de bloedvaten. Dit is het gevaarlijkst als bijvoorbeeld de bloedvaten van de hersenen en het hart dichtslibben. Doordat de doorgang in de bloedvaten nauwer is geworden, kan het bloed moeilijker het gebied achter de vernauwing bereiken. Bij vaatvernauwing kan achterliggend weefsel te weinig zuurstof krijgen en afsterven. Dit kan in een ernstig geval leiden tot het sterven van een organisme. Over het algemeen werd er in het verleden gedacht dat pathologische mineralisatie het gevolg was van fysiochemische precipitatie (geen cellen voor nodig) van een te hoge calcium en fosfaat concentratie. Recente onderzoeken hebben echter aangetoond dat pathologische mineralisatie een gereguleerd proces is dat veel gelijkenissen heeft met het gereguleerde proces van botvorming. In een organisme wordt tijdens de embryogenese 2 anatomisch verschillende typen bot gevormd: pijpbeenderen (bv. een dijbeen) en platte beenderen (bv. de schedel). Deze typen ontstaan op verschillende wijzen. Pijpbeenderen ontwikkelen zich door een proces van botvorming uit kraakbeen: endochondrale ossificatie. Tijdens de endochondrale ossificatie differentiëren mesenchymale stamcellen (een type stamcel uit het beenmerg) zich op een bepaald moment in kraakbeencellen (chondrocyten) (Hfst 1, Fig. 1, panel 2). Deze cellen zijn ingebouwd in een collageen matrix die voornamelijk bestaan uit water, suikers en type-II collageen en vormen door middel van celdeling het eerste been model. Op een bepaald moment gaan de chondrocyten mineraliseren. Vervolgens invaseren bloedvaten het gemineraliseerde kraakbeen. Via deze bloedvaten migreren botcellen (osteoblasten) naar het gemineraliseerde kraakbeen. Osteblasten groeien dan op de gemineraliseerde matrix en zetten een botmatrix af dat voornamelijk bestaat uit Type-I collageen. De chondrocyten verdwijnen voor een groot deel, maar in een specifiek gebied van het gevormde bot, de groeischijf, blijven de chondrocyten bestaan en delen. Aan de uiteinden van de groeischijf mineraliseren chondrocyten en uit het bloed dringen vervolgens weer osteoblasten het beginnende botstuk


123

binnen die groeien op de gemineraliseerde matrix. In het centrum van de groeischijf blijven chondrocyten delen (zo word je langer). Dit proces gaat door tot het einde van de puberteit. Dan mineraliseert de groeischijf volledig. Platte beenderen ontstaan door zogenaamd intramembraan ossificatie (Hfst 1, Fig. 1, panel 1). Hierbij is er geen kraakbenige tussenfase, maar vormen de mesenchymale stamcellen direct losmazig collageen. Osteoblasten komen ingebed te liggen in het gevormde, gemineraliseerde netwerk en vormen vervolgens een nieuwe botmatrix. Hieruit ontstaat door remodellering een georganiseerde botstructuur, net zoals bij endochondrale ossificatie. Osteoblasten zijn gespecialiseerde stromacellen die exclusief zorgen voor de vorming, de afzetting en de mineralisatie van botweefsel. Ze komen voort uit dezelfde basiscellen als de chondrocyten, namelijk de mesenchymale stamcellen. De differentiatie van de stamcellen wordt gereguleerd door activering van specifieke transcriptiefactoren (DNA regulatoren) en eiwitten. Wanneer de differentiatie van osteoblasten uit mesenchymale cellen eenmaal in gang is gezet, ontwikkelen de zogenoemde osteoprogenitoren, snel tot pre-osteoblasten. Als rijpe osteoblasten de botmatrix hebben gevormd, gaat een deel dood (ongeveer 70%) en wordt een deel (ongeveer 30%) ingevangen in de botmatrix. Deze ingevangen cellen differentiëren tot osteocyten. Osteocyten kunnen geen bot vormen of afbreken. Hun taak is de architectuur van het bot in stand te houden en ervoor te zorgen dat bot zich kan aanpassen aan mechanische en chemische invloeden. Naast de aanmaak van bot door osteoblasten, wordt bot ook continu afgebroken door gespecialiseerde cellen, de osteoclasten. Osteoclasten worden gevormd door fusie van mononucleaire voorlopercellen die voortkomen uit hematopoïetische (bloedvormende) stamcellen. De regulering van de botafbraak door osteoclasten omvat een complex geheel van interacties tussen hormonen, osteoblasten, osteocyten en osteoclasten. Tijdens groei zijn voornamelijk osteoblasten actief (opbouw van bot) en tijdens veroudering zijn de osteoclasten (afbraak van bot) voornamelijk actief. Recent onderzoek heeft aangetoond dat tijdens pathologische mineralisatie gladde spiercellen uit bijvoorbeeld de bloedvaten kunnen differentiëren tot osteoblasten. Tijdens aderverkalking wordt soms compleet bot met osteoblasten, osteoclasten en beenmerg gevormd. Hetzelfde is inmiddels bekend van gemineraliseerd articulair gewricht kraakbeen (artrose). Gladde spiercellen zijn net als botcellen en kraakbeencellen afkomstig van de mesenchymale cellen. Het zou mogelijk kunnen zijn dat om de één of andere reden deze cellen differentiëren tot botcellen, dat theoretisch een eindstadium lijkt te zijn van mesenchymale cel differentiatie. Het is echter tot op heden niet bekend waarom deze cellen mineraliseren en hoe dit mechanisme gereguleerd is.


124

Resultaten beschreven in dit proefschrift In het eerste hoofdstuk wordt beschreven wat er in de literatuur bekend is over cel-gemediëerde mineralisatie. In dit hoofdstuk worden 4 verschillende theorieën beschreven over hoe mogelijk cel-gemediëerde mineralisatie is gereguleerd. De eerste theorie omschrijft de rol van matrix vesicless bij het begin proces van mineralisatie. Matrix vesicles zijn kleine blaasjes (ongeveer 0.1-1 µm groot) die afgesnoerd worden van de plasmamembraan van de chondrocyt, osteoblast of gladde spiercel. Wat nu precies de cel stimuleert om deze blaasjes af te snoeren is niet bekend. Het is wel bekend dat nadat deze blaasjes afgesnoerd zijn en geïmmobiliseerd zijn in de extracellulaire matrix, ze calcium en fosfaat opnemen en deze ionen precipiteren. Op deze manier zou mogelijk het eerste kristal kunnen ontstaan tijdens het mineralisatie proces. De tweede theorie omschrijft de rol van geprogrammeerde celdood (apoptose) bij het beginproces van mineralisatie. Een verhoogd fosfaat concentratie zou ervoor kunnen zorgen dat geprogrammeerde celdood geïnduceerd wordt. Tijdens geprogrammeerde celdood splitsen kleine bollen met cytoplasma met celinhoud zich af (apoptotische lichaampjes) waarin calcium- en fosfaat ionen mogelijk kunnen precipiteren. Dit gebeurt waarschijnlijk omdat in deze apoptotische lichaampjes specifieke negatief geladen vetten (phosphatidylserine) op het oppervlak van de plasmamembraan voorkomen. Deze vetten binden calcium en samen met een verhoogt concentratie fosfaat in de weefselvloeistof precipiteren deze ionen tot een kristal. De derde theorie omschrijft de rol van specifieke eiwitten bij het begin proces van mineralisatie. De toegenomen fosfaat concentratie in de weefselvloeistof zorgt ervoor dat de cel bepaalde eiwitten gaat uitgescheiden in de extracellulaire matrix. Deze eiwitten kunnen calcium- en fosfaat ionen precipiteren. Daarnaast kunnen specifieke eiwitten ook mineralisatie remmen. Als één van deze eiwitten verminderd of afwezig is, ontstaat mineralisatie. In serum (bloedcomponenten) zit bijvoorbeeld een hele belangrijke sterke mineralisatie remmer. Cellen produceren zelf ook mineralisatie remmers. Samengevat staan de verbanden tussen deze verschillende theorieën in Hfst 1, Fig. 2. Gebaseerd op de verschillende theorieën zijn er nog veel vragen over hoe nu precies cel gemediëerde mineralisatie begint. Daarom was de doelstelling van dit promotieonderzoek, te bestuderen hoe mineralisatie begint in een celkweeksysteem. Daarvoor is gebruik gemaakt van de ATDC5 cellijn. Dit is een muizen cellijn die geschikt is om mineralisatie op cel niveau te bestuderen. In 2 weken ontwikkelen deze cellen een extracellulaire matrix en kan mineralisatie geïnduceerd worden na toevoeging van extra fosfaat. Mineralisatie is dan na 24 uur te detecteren. De resultaten beschreven in het hoofdstuk 2 laten zien dat het verwijderen van serum in het celkweek system resulteert in de formatie van grote mineralisatie


125

structuren. Dit proces duurt 2 uur. Verder is waargenomen dat in het celkweeksysteem zonder serum de grote mineralisatie structuren drijven in het celkweek medium (vloeistof op de cellen) en ze zichtbaar zijn met het blote oog. Tijdens microscopische studies van de grote mineralisatie structuren is waargenomen dat de cellen in deze structuren uit elkaar getrokken zijn. Een mogelijke verklaring hiervoor is dat het eerste calcium fosfaat kristal intracellulair (in de cel) ontstaat. In Fig. 7 van hoofdstuk 2 en appendix A bij hoofdstuk 2 is daarom bestudeerd waar de mineralisatie structuren met de mineralisatie remmer, serum, zich bevinden in het celkweeksysteem. Uit microscopische studies blijkt dat de mineralisatie structuren in het celkweeksysteem met serum klein zijn (~ 0.1 – 1 µm) en dat de mineralisatie structuren nog intracellulair liggen in cellen die geprogrammeerd zijn om dood te gaan. Om uit te sluiten dat de formatie van de grote mineralisatie structuren niet geïnduceerd wordt door enkele dode (apoptotische) cellen in het medium van het celkweeksysteem, werden cellen vooraf doodgemaakt en zonder serum geïnduceerd om te mineraliseren. Echter, vooraf doodgemaakte cellen zonder serum kunnen geen grote mineralisatie structuren vormen. Dit wijst erop dat een cel in eerste instantie levend moet zijn om grote mineralisatie structuren te vormen. Verder blijkt uit de resultaten van het hoofdstuk 2 dat een fosfaat transporter remmer (remt fosfaat transporters in de membraan) de formatie van de grote mineralisatie structuren remt. Dit suggereert dat een cel eerst fosfaat moet opnemen om het mineralisatie proces te starten. Vervolgens is onderzocht of fosfaat mineralisatie in de cel (intracellulair) induceert. Daarvoor is gekeken of annexin V transloceert van het cytosol naar de membraan. Annexin V is een eiwit dat zich normaal gesproken bevindt in het cytosol, maar in aanwezigheid van een verhoogd calcium concentratie in de cel aan de cel membraan bindt. Uit de resultaten van dit onderzoek blijkt dat fosfaat annexin V translocatie induceert van het cytosol naar de membraan in alle cellen in het celkweek systeem. Waarschijnlijk gebeurt dit doordat fosfaat de cel stimuleert intracellulair de calcium concentratie te verhogen. Daarom is in appendix B bij hoofdstuk 2 de calcium concentratie gemeten in het cytosol van cellen die geïnduceerd waren om te mineraliseren. De calcium concentratie nam in het cytosol niet toe in de mineralisatie geïnduceerde cellen. Mogelijk induceert fosfaat de annexin V translocatie via een ander mechanisme of wordt calcium direct door de cellen weggepompt in specifieke organellen (bijvoorbeeld het endoplasmatisch reticulum of de mitochondriën). In hoofdstuk 3 is bestudeerd of de ATDC5 cellen eiwitten afgeven die de formatie van calcium fosfaat kristallen beïnvloeden. Uit de resultaten van dit hoofdstuk blijkt dat cellen continu oplosbare factoren afgeven die de formatie van calcium fosfaat kristallen kunnen beïnvloeden. Deze eiwitten zijn geïsoleerd uit de kristallen en gekarakteriseerd. Uit de proteomics data (eiwit analyse) blijkt dat er


126

een specifieke set van serum eiwitten, cytosolaire eiwitten en extracellulaire matrix eiwitten inzitten. Bulk cytosolair eiwitten afkomstig van gehomogeniseerde cellen en serum eiwitten konden niet de formatie van calcium fosfaat kristallen beïnvloeden. Verder is waargenomen dat fibronectin knock-out cellen (cellen die het matrix eiwit fibronectin niet meer kunnen vormen) nog steeds de vorming van calcium fosfaat kristal kunnen beïnvloeden, net zoals cellen die wel fibronectin kunnen vormen. Uit de resultaten van dit hoofdstuk blijkt dat de cellen specifieke nog niet gekarakteriseerde oplosbare factoren afgeven die de vorming van het kristal kunnen beïnvloeden. In hoofdstuk 4 is het effect van de stikstof mono-oxide (NO) donor sodium nitroprusside (SNP) op mineralisatie bestudeerd. SNP is een farmacologische stof dat veel gebruikt wordt bij harttransplantaties om de bloeddruk te verlagen. Uit de resultaten van dit hoofdstuk blijkt de SNP bij een zeer lage concentratie (10 µM) mineralisatie remt in het celkweeksysteem. De relevantie van dit onderzoek is dat SNP naast het verlagen van de bloeddruk mogelijk ook pathologische mineralisatie remt. Deze remming bleek niet via het NO/cyclisch-guanosinemonofosaat (molecuul) gemediëerde signaal transductie route (een signaal dat wordt herkend door de cel en vervolgens omgezet in een specifieke set van gebeurtenissen in de cel) mineralisatie te remmen. SNP bleek noch mineralisatie te remmen door fosfaat transport in de cel te remmen en noch door het remmen van geprogrammeerde celdood. In hoofdstuk 6 is het effect van SNP op mineralisatie verder onderzocht. Uit de resultaten van dit hoofdstuk blijkt dat SNP niet mineralisatie remt door NO af te geven, maar door uit elkaar te vallen in een ijzer component. IJzer blijkt net als SNP bij een zeer lage concentratie mineralisatie in het celkweeksysteem te remmen.

Abbreviations APase alkaline phosphatase β-GP betaglycerophosphate BAG-75 bone acidic glycoprotein-75 BSA bovine serum albumin BSP bone sialoprotein cGMP cyclic guanosine-3’,5’monophosphate cGK cGMP dependent protein kinases () DMP1 dentin matrix protein-1 EDTA ethyleendiamine tetraacetate FMS flat mineralizing structures FT-IR fourier transform infrared HAP hydroxyapatite LMS large mineralizing structures MVs matrix vesicles NO nitric oxide NOS nitric oxide synthase NCPs non-collagenous proteins OC osteocalcin ON osteonectin OPN osteopontin PBS phosphate buffered saline Pi phosphate PS phosphatidylserine PFA phosphonoformic acid SDS PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis sGC soluble guanylyl cyclase SNP sodium nitroprusside

127

129

Dankwoord Zo en dan nu het dankwoord. Natuurlijk is dit proefschrift niet door mij alleen tot stand gekomen: daar zijn er vele mensen bij betrokken geweest. Allereerst wil ik mijn co-promotor Bas Vaandrager bedanken. Bas, wat ben ik blij dat ik jou als begeleider heb gekregen. Ik heb zoveel van je geleerd en dat alles op zo’n prettige manier. Altijd stond je deur open, zodat ik de vrijheid nam regelmatig binnen te lopen als ik iets wilde overleggen of het gewoon niet meer snapte. Ik zal de samenwerking met jou en onze vele wetenschappelijke discussies missen. Ook jouw enthousiasme voor de wetenschap heeft aanstekelijk gewerkt en als ik het even niet meer zag zitten, dacht ik ‘gewoon even met Bas kletsen’ en dan zag ik het daarna allemaal weer positief in. Verder wil ik mijn promotor Bernd Helms en co-promotor René van Weeren bedanken. Elke maand hadden we een gezamenlijke werkbespreking. Tijdens die werkbesprekingen was de sfeer ontspannen, gezellig, leerzaam en werden er vooral nieuwe ideeën gecreëerd. Naar mijn mening vormden we met zijn vieren een goed team waarin ieder zijn eigen kwaliteit had. Mijn promotor Ab Barneveld wil ik ook bedanken. Ik heb je gedurende het project niet vaak gezien, maar ik begreep later dat je op de achtergrond alles nauwlettend in de gaten hielt. Als team van begeleiders wil ik jullie tenslotte bedanken voor de grote mate van vrijheid en voor het vertrouwen dat ik van jullie heb gekregen. Dit heeft zeker een positieve bijdrage geleverd aan mijn wetenschappelijke ontwikkeling. Verder wil ik de mensen bedanken die mij met het praktische werk geholpen hebben. Ik herinner me nog goed dat toen ik aan mijn promotieonderzoek begon, ik nog nooit een celkweek fles had gezien. Wil Klein en Martin Houweling, ik wil jullie bedanken dat jullie mij de fijne kneepjes van deze labvaardigheid hebben geleerd. Mijn project was er immers honderd procent afhankelijk van. Voor de hulp bij de calcium bepaling wil ik Hugo Wouterse bedanken. Ik herinner me nog de eerste keer dat ik met behulp van deze meting na een jaar zweet, bloed en tranen voor het eerst mineralisatie in het celkweeksysteem waarnam. Dat was een mooi moment. Zoals blijkt uit mijn proefschrift heb ik ook heel veel gebruik gemaakt van de faciliteiten van het centrum voor cellulaire beeldvorming. Voor het leren omgaan met de elektronen microscoop, fluorescentie microscoop en de multifoton microscoop wil ik Ton Ultee, Laura van Weeren en Anko de Graaff bedanken. Ik heb er bewondering voor met hoeveel geduld jullie de bediening van deze apparatuur kunnen uitleggen en jullie leken ook nooit geïrriteerd als ik in het begin jullie binnen 5 minuten tenminste 3 keer moest storen. Dan had ik weer op een

130

knopje gedrukt waarvan ik de uitwerking niet begreep. Of het apparaat deed simpelweg gewoon niet wat ik wilde. Als dan het apparaat toch deed wat het moest doen en ik toch niet het bevredigende resultaat zag, dan ging ik naar Richard Wubbolts. Richard, bedankt voor je tijd, uitleg en goede tips. Aan het einde van mijn project heb ik van Remco Westerink nog een techniek geleerd voor intracellulair calcium imaging. Remco, ik heb de samenwerking met jou als prettig ervaren. Ook wil ik alle collega’s van de hoofdafdeling biochemie en celbiologie bedanken voor de gezelligheid en roddels op het lab. Het was een interessante werkomgeving. Renske, Edita, Yi-lo en Patricia, ik heb leuke gesprekken met jullie gehad en jullie waren gezellige labgenoten. Ruud, Ynske, Josse en Mijke, ik heb hele leuke tijden met jullie op het lab gehad en ik wil nog even kwijt dat meer dan je denkt om de pH draait. Michiel, Suus, Anita, Koen v G. jullie waren gezellige kamergenoten. Verder wil ik ook Kim, Onno, Klaas-Jan, Martin v E., Danny, Evert, Karin, Jan-Willen en Sonja bedanken voor de gezelligheid op de afdeling. Wat ik vooral prettig aan jullie allemaal heb gevonden is dat er altijd wel iemand was waarbij je kon klagen of zeuren als je daar behoefte aan had. Meestal bleek je dan niet de enige te zijn en gedeelde smart is halve smart. De humor die daaruit voortkwam vond ik echt geweldig. Voor de feedback, betrokkenheid en interesse wil ik ook Dora, Jaap, Bart, Jos, Coos, Lodewijk en Marion bedanken. Judith, jou wil ik ook nog bedanken voor het regelen van praktische zaakjes zoals: waar kan ik het beste subsidies aanvragen, ik wil een nieuwe computer, mijn stoel is verrot, wanneer heeft Bernd nou eens tijd, wie moet ik hebben van personeelszaken, waar staan nu de huidige promotie reglementen op de website en hoe werkt nu dat nieuwe systeem enz, enz. Je handelde dit soort zaakjes altijd verbazingwekkend doortastend en snel af. Harry Otter van de afdeling Multimedia wil ik bedanken voor de lay-out van mijn proefschrift. Dat deed je binnen een paar uur en waardoor ik veel sneller klaar was met het proefschrift dan ik van te voren gedacht had. Naast de mensen van het lab ben ik ook dank verschuldigd aan andere mensen in mijn omgeving. De uitslag van mijn experimenten had een behoorlijk effect op mijn humeur, zeker de eerste paar jaar. Voor de steun in deze perioden wil ik mijn paranimfen Lot en Thirza bedanken. Ik heb het als heel prettig ervaren dat jullie altijd klaar voor me staan en achter mij staan. Als ik met één van jullie gesproken had, was de donkere wolk die boven mijn hoofd hing weer weggewaaid. Ook wil ik Linda, Petra, Ingrid, Leonie B., Marieke van de V. en mijn vrienden die ik ontmoet heb in Nieuw Zeeland (Ariëtte, Marieke K, Renske v T., Tamara, Janneke, Erik, Dennis en Evert) bedanken voor een luisterend oor. Lieve Koen hoe ik jou moet bedanken zou ik gewoon niet weten. Allereerst was je in mijn eerste jaar een gezellige humoristische collega, waar vooral mee te lachen

131

viel. Dat werd al gauw een interessante collega waar meer in zat. Daarna heb ik zoveel steun van je gehad. Zowel inhoudelijke ideeën voor het project dat zeer nuttig was als ook een luisterend oor wanneer er tegenslag was in het project. Een kunst van jou is, dat wanneer mijn humeur echt een diepte punt heeft bereikt, je me toch nog aan het lachen kunt maken. Dan nog mijn broer en lieve ouders. Wat ben ik blij met jullie en wat hebben jullie mij goed gesteund. Altijd staat bij jullie de deur open, is het er warm, gezellig en vertrouwd. Dat geeft toch zo’n prettig geborgen gevoel. Zodat wanneer ik bij julie ben, me realiseer dat alles maar relatief is. Leonie

132

Curriculum Vitae De auteur van dit proefschrift werd geboren op 30 juli 1978 te Soest. De HAVO en daaropvolgend het VWO werden doorlopen op het Baarnsch Lyceum te Baarn, waar in 1995 het HAVO diploma werd behaald en in 1997 het VWO diploma. In augustus datzelfde jaar werd begonnen met de studie Medische Biologie aan de Vrije Universiteit van Amsterdam. Tijdens de studie Medische Biologie voerde zij haar eerste stage uit bij de hoofdafdeling Gezelschapsdieren in samenwerking met de hoofdafdeling Gezondheidszorg Paard aan de Faculteit der Diergeneeskunde aan de Universiteit Utrecht onder begeleiding van Prof. Dr. B. van Oost en Dr. W. Back. Het project betrof genetisch onderzoek naar osteochondrose bij het KWPN paard en onderzoek naar de overerving van de vachtkleur bij het Nederlands Trekpaard. Haar tweede stage voerde zij uit aan de Faculty of Veterinary Medicine aan de Massey University in Nieuw Zeeland onder begeleiding van Prof. Dr. E. Firth en Dr. C. Rogers. Het project betrof de kwantificatie van de impact op het musculoskelet systeem tijdens de training van Engels Volbloed renpaarden. Na het afronden van deze stage werd het docteraal diploma in 2002 gehaald. Direct aansluitend werd een aanstelling verkregen als Assistent in Opleiding (AIO) bij de hoofdafdeling Biochemie en Celbiologie in samenwerking met de hoofdafdeling Gezondheidszorg Paard aan de Faculteit der Diergeneeskunde aan de Universiteit Utrecht. Tijdens deze aanstelling is promotieonderzoek verricht onder begeleiding van Prof. Dr. J.B. Helms, Prof. Dr. A. Barneveld, Dr. A.B. Vaandrager, Dr. P.R. van Weeren, Dr. C.H.A. van de Lest.

Characterization of biological mineralization in vitro

Documents

Transcript of Characterization of biological mineralization in vitro