|Home | About | Journals | Submit | Contact Us | Français|
The presence of bone-marrow-derived stem cells was investigated in a wound-healing model where subcutaneously implanted cellulose sponges were used to induce granulation tissue formation. When cellulose was coated with hydroxyapatite (HA), the sponges attracted circulating haemopoietic and mesenchymal progenitor cells more efficiently than uncoated cellulose. We hypothesized that the giant cells/macrophages of HA-coated sponges recognize HA as foreign material, phagocyte or hydrolyse it and release calcium ions, which are recognized by the calcium-sensing receptors (CaRs) expressed on many cells including haemopoietic progenitors. Our results showed, indeed, that the HA-coated sponges contained more CaR-positive cells than untreated sponges. The stem cells are, most probably, responsible for the richly vascularized granulation tissue formed in HA-coated sponges. This cell-guiding property of HA-coated cellulose might be useful in clinical situations involving impaired wound repair.
Wound healing is a complex process involving the coordinated interaction of a variety of tissue and cell types. Numerous studies suggest that bone-marrow-derived stem cells participate in tissue repair, e.g. in cutaneous wound healing (Wu et al. 2007). Traditionally, bone marrow has been thought to consist of two distinct cell types, namely haemopoietic stem cells (HSCs) and bone marrow stromal cells (MSCs). HSCs differentiate into all mature blood cells whereas MSCs act as a supportive scaffold and a rich source of growth factors to maintain the ideal microenvironment for cells in the bone marrow. During the last few years, the conception of bone marrow stem cells has changed enormously. HSCs are now considered as a heterogenic cell population that contains not only stem cells committed to haemopoiesis but also endothelial progenitor cells (EPCs; Asahara et al. 1999).
Tissue ischaemia and cellular damage caused by trauma release cytokines, which mobilize EPCs from the bone marrow into peripheral blood, and the circulating EPCs then home to sites of nascent neovascularization and differentiate into mature vascular endothelial cells (Takahashi et al. 1999; Kawamoto et al. 2003). Besides haemopoietic and vasculogenic plasticity, adult HSCs/EPCs have been shown to possess an ability, at least in vitro, to differentiate into non-blood cells, for example to osteoblasts (Long et al. 1990; Chen et al. 1997; Tondreau et al. 2005; Matsumoto et al. 2008), cardiomyocytes (Badorff et al. 2003; Zhang et al. 2004; Iwasaki et al. 2006) and even to skeletal, hepatic and neural lineages (Ratajczak et al. 2004).
Bone-marrow-derived MSCs are multipotent cells capable of differentiating into numerous cell types, e.g. fibroblasts, cartilage, bone, muscle, brain cells (Sordi et al. 2005; Caplan & Dennis 2006; Valtieri & Sorrentino 2008). The MSCs also secrete a large number of growth factors and cytokines that are critical to the repair of injured tissues (Caplan & Dennis 2006). In addition to their multipotency, MSCs have also immunosuppressive effects (Koç et al. 2000; Caplan & Dennis 2006; Sasaki et al. 2008). Owing to the plasticity of the MSCs, they are widely used in experiments to improve cutaneous wound healing and the results are promising (Yamaguchi et al. 2005; Liu et al. 2006; McFarlin et al. 2006; Cha & Falanga 2007; Kwon et al. 2008).
Wound healing is characterized by the formation of a richly vascularized granulation tissue, which supports the increased nutritional needs of rapidly proliferating and migrating inflammatory and progenitor cells. Implantation of cellulose sponges offers a convenient model to study granulation tissue formation experimentally. The material does not interfere with the normal wound repair process but acts as a chemoattractant for cells that are involved in the process. Cellulose has been used to induce granulation tissue growth not only for research purposes (Märtson et al. 1999; Inkinen et al. 2003) but also in clinical studies of wound healing as well (Gay et al. 1978; Viljanto et al. 1981; Pajulo et al. 1999, 2001). In our previous work, we showed that hydroxyapatite (HA) coating of cellulose sponges induces a more effective formation of a highly vascularized granulation tissue than uncoated cellulose when the sponges were implanted subcutaneously in rat (Tommila et al. 2008). The HA coating was formed by treating cellulose with a bioactive glass, which creates a silica-rich calcium phosphate layer (Ekholm et al. 2005). In the bone marrow, the HSC niche is located close to the endosteal surface and is regulated via calcium-sensing receptors (CaRs) of HSCs (Adams et al. 2006; Drüeke 2006). The HA coating resembles the natural mineral composition of bone (Ekholm et al. 2005) and might therefore attract bone-marrow-derived stem cells. Thus, we wanted to investigate whether the granulation tissue-stimulating ability of the HA-coated cellulose implant is due to increased homing of bone-marrow-derived stem cells.
HA-coated cellulose was prepared as described previously (Ekholm et al. 2005). In short, simulated body fluids (SBF) with ion concentration close to 1 and 1.5 times that of human blood plasma were prepared, buffered to pH 7.4 and sterilized by filtering. The initial calcium phosphate layer was precipitated on regenerated, oxidized cellulose sponges (Vivoxid/Cellomeda Ltd, Finland) by the biomimetic method of Kokubo et al. (1991). Mineralization was initiated in 500ml 1×SBF supplemented with 2g of a specific granular bioactive glass (S53P4, Vivoxid Ltd) at 37°C for 24 hours and was then grown in 500ml 1.5×SBF at the same temperature under continuous shaking. After 14 days, the sponges were washed briefly with sterile water, dried at room temperature in a laminar hood and the presence and homogeneity of the mineral layer were confirmed by scanning electron microscopy.
Mineralized and untreated celluloses were cut into pieces of 5×5×10mm, autoclaved and implanted into nine male Sprague–Dawley rats (mean weight 304±17g, age 10–13 weeks). The animals were anaesthetized with a subcutaneous injection of a mixture of midazolam (Dormicum, Roche, Switzerland) and fentanyl–fluanisole (Hypnorm, Jansen Pharmaceuticals, Belgium) in sterile water (1:1:2). The backs of the animals were shaved, disinfected with a 5mgml−1 solution of chlorhexidine gluconate (Klorhexol, Leiras, Finland) and draped with sterile towels. Two separate 2cm long midline incisions, cranial and caudal, were made on the back of the animals. The sterilized sponge implants were moistened with 0.9 per cent NaCl solution and inserted bilaterally into the pockets of the subcutaneous space without fixation. Mineralized and non-mineralized implants were inserted alternately into to the left and right sides to reduce biological variation. Skin wounds were closed with interrupted absorbable sutures (4-0 Vicryl-rapid, Ethicon).
The rats were allowed to recover from anaesthesia in the operation room under a heating lamp and were housed individually in the cages with free access to food pellets and water. The animals were killed via cervical dislocation under CO2 anaesthesia: three animals at day 3 and six animals at day 7 post-operatively. The implants were dissected free from the surrounding connective tissue. All implants from day 3 and the implants from three animals from day 7 were fixed in 4 per cent phosphate buffered paraformaldehyde for 16 hours for histology. The implants from the remaining three rats at day 7 were used in cell-sorting analysis.
The animal facilities are managed according to the European Convention for the Protection of Vertebrate Animals used for Experimental and Other Scientific Purposes and inspected for GLP Compliance. The study design was approved by the Animals Ethical Committee of the State Provincial Office of Southern Finland.
The fixed sponge samples were embedded in paraffin and sectioned longitudinally with a microtome. The sections were stained with haematoxylin–eosin (HE) and Giemsa for light microscopy examination.
Immunostainings of formalin-fixed paraffin-embedded sections were carried out as described previously (Tommila et al. 2008) using the avidin–biotin–peroxidase complex technique (Vectastain ABC kit, Vector Laboratories Inc., USA). The antibodies used in this work as stem cell markers were CD34 (rat monoclonal; Santa Cruz Biotechnology, USA), c-kit (CD117; rabbit polyclonal; Santa Cruz Biotechnology) and stro-1 (mouse monoclonal; Zymed Laboratories, USA). All primary antibodies were diluted 1:50. CD34 is the most commonly used HSC marker, whereas c-kit is generally found in many premature cell lineages. Stro-1, notably, has a superior position as a mesenchymal stem cell marker. For counter-staining, the sections were kept in Papanicolaou's solution 1a, Harris' haematoxylin solution (Merck, Germany) for 30s.
CaR was visualized by immunofluorescence using an anti-CaR mouse monoclonal antibody (1:50; Santa Cruz Biotechnology). Texas-Red-labelled goat anti-mouse IgG (Molecular Probes, USA) was used as a secondary antibody. Nuclei were counter-stained with 4′, 6-diamidino-2-phenylindole (Molecular Probes).
Cells from 7-day-old implants of three rats were detached enzymatically. Immediately after removal, the implants were placed on ice in Falcon tubes with 2ml of α-minimum Eagle's medium (α-MEM; Gibco, USA) complemented with 15 per cent foetal bovine serum (Biochrom AB, Germany). The implants were cut into small pieces on a Petri dish in a laminar flow hood and placed into new Falcon tubes, and 2ml of prewarmed (37°C) enzyme mixture (0.1% collagenase (Boehringer Mannheim, Germany) and 0.05% trypsin (Sigma, USA) in sterile phosphate-buffered saline) was pipetted on top. The tubes were gently shaken for 45min at 37°C, and then the samples were filtered through nylon membranes with 2ml of α-MEM. The cell mixture was centrifugated (1600g, 4min) and the supernatant was removed. The cells were washed twice with 2ml of α-MEM and finally suspended in 1ml of α-MEM and used for flow cytometry.
Flow cytometric measurements were done with a FACSCalibur instrument (Becton Dickinson, USA) and data analysis was performed with Cyflogic software (Cyflo, Finland). The living cells were gated from forward-scatter/side-scatter dot plot. The proportion of c-kit-positive cells was measured from the samples. Alexa Fluor 488-conjugated goat anti-rabbit IgG (Molecular Probe) was used as the secondary antibody to detect anti-c-kit (CD117; rabbit polyclonal, Santa Cruz Biotechnology).
The mean fluorescence intensity of HA-coated samples versus uncoated controls was compared using two-tailed t-test. Values are presented as mean±s.d. A value of p<0.05 was considered significant.
As we reported earlier (Tommila et al. 2008), granulation tissue grew notably faster into the HA sponges compared with uncoated sponges and filled the HA-coated sponges within two weeks. This phenomenon was also evident in the Giemsa-stained sections, in which the staining was more intense in HA sponges both at days 3 and 7 (figure 1 a–c). The foreign body reaction started from the sponge surfaces from where the cells invaded further towards the sponge centres. The cells in the untreated control sponges appeared similar to the cells in the HA-coated sponges, but were in smaller quantity (figure 1 d).
Histological stainings of cellulose implants show more cells in the HA-coated sponges. The invading cells are gathered around the cellulose fragments (cf): (i) coated, (ii) uncoated. Giemsa-stained (a) 3- and (b) 7-day-old implants (scale bar, 25μm). ...
Enzymatically detached cells from 7-day-old implants were screened for the expression of c-kit, a common premature cell marker. There was a significantly (p<0.05) higher c-kit expression in HA-coated sponges than in the uncoated ones (figure 2).
Flow cytometry analysis revealed more c-kit-positive cells in HA-coated 7-day-old implants. (a) Histogram showing the expression of c-kit-positive cells extracted from the cellulose implants. The filled histogram under the segment of a line shows the ...
All stem cell markers used in this study stained numerous cells in the HA sponges, whereas hardly any positive cells were observed in the controls (figure 3). The stainings of c-kit and CD34 were emphasized in the sponge centre, while the common mesenchymal stem cell marker, stro-1, stained cells closer to the sponge surfaces, i.e. the area that was rich in collagen-containing granulation tissue. All studied stem cells, especially in the HA-coated implants, were located in close contact with cellulose fragments.
Stem cell markers in 7-day-old implants ((i) coated and (ii) uncoated). (a) Numerous c-kit-positive cells were observed in the sponge centre of HA-coated implants. (b) Anti-CD34-stained small rounded cells in the middle of the HA-coated sponge. (c) Stro-1, ...
The CaR is expressed in many cell types, including macrophages and monocytes. These receptors have also proven to be important in the HSC storage in bone marrow (Drüeke 2006). An abundant amount of CaR-positive cells was observed in HA-coated samples not only in the new granulation tissue, but also in the central area of the implants. Most of the CaR-positive cells were gathered closely to the HA-coated cellulose fragments (figure 4). Similar cells were seen in the plain cellulose sponges, but in remarkably lesser amounts and mainly near the implant surfaces.
Calcium-sensing receptor-positive cells in 7-day-old implants. (a) CaR-positive cells in the centre of the HA-coated sponge are accumulated in clusters near cellulose fragments. (b) Only a few CaR-positive cells can be seen in the control. (i) Lower magnification ...
At a wound site, local and infiltrated cells release chemokines, which recruit blood-circulating stem and progenitor cells. These chemokines also increase bone marrow cell mobility, thus facilitating cell mobilization into the peripheral blood and into the sites of wound healing (Sasaki et al. 2008). When implanted subcutaneously in the rat, the HA-coated cellulose sponges induced a stronger and quicker inflammatory response than untreated sponges. The HA layer augmented the recruitment of macrophages, which provide a continuing source of growth factors and cytokines that affect the local environment (Tommila et al. 2008).
In the present study, we showed that the HA coating of cellulose sponges also increased the homing of bone-marrow-derived stem cells. The implants were screened for c-kit expression by flow cytometry, and a significantly higher expression was observed in HA-coated samples. Different types of stem cells were further identified by immunohistochemistry. C-kit- and CD34-positive cells indicated that HSCs were the first ones to invade the empty centres of the sponges, while MSCs (stro-1-positive) were found in the forming granulation tissue.
In the bone marrow, undifferentiated HSCs are detected near the endosteum, the thin layer of cell-rich connective tissue that lines the medullary cavity of long bones, in the so-called endosteal stem cell niche (Haylock & Nilsson 2005). At this site, the bone is in a constant turnover; bone is laid down by osteoblasts and removed by specific macrophages, the osteoclasts. Owing to bone degradation, soluble calcium ions (Ca2+) are released in the bone marrow fluid. Cells, including haemopoietic progenitors as well as monocytes/macrophages, respond to extracellular ionic calcium concentrations through a calcium-sensing G-protein-coupled receptor, CaR. Therefore, CaR seems to have a function of retaining HSCs in close physical nearness to the endosteal surface (Adams et al. 2006). The HA coating used in this study resembles the HA of bone (Ekholm et al. 2005). Plentiful giant cells/macrophages are gathered around the mineralized cellulose (Tommila et al. 2008). We hypothesized that while macrophages try to phagocyte or hydrolyse the silica-rich foreign material, Ca2+ is released, which creates a beneficial milieu for the gathering of circulating haemopoietic progenitors. Indeed, CaR-positive primitive cells were observed near the mineralized cellulose fibres, supporting our theory that circulating haemopoietic progenitors home to the mineralized cellulose and differentiate into various haemopoietic cells.
According to our microarray data of an ongoing study to further characterize gene expression in granulation tissue (unpublished data), cells in the granulation tissue in the cellulose sponges express stromal-derived factor-1 (SDF-1), a vital chemokine for stem and progenitor cell recruitment in tissue repair after injury (Kollet et al. 2003; Ratajczak et al. 2004; Kaplan et al. 2007; Schantz et al. 2007). The microarray data also revealed that the expression of CXCR4, the receptor for SDF-1, is significantly upregulated in HA-coated sponges, indicating more stem and progenitor cells in these sponges. The homing mechanisms that guide the recruitment of MSCs are not yet fully understood, but SDF-1/CXCR4 is described to be an important part of the homing process (Wynn et al. 2004; Schantz et al. 2007).
In the bone marrow microenvironment, there seems to be an important functional dialogue between HSCs/EPSs and MSCs (Satoh et al. 1997; Valtieri & Sorrentino 2008), but their exact roles in tissue repair remain unclear. It is likely that both cell types are needed in wound healing. MSCs express transcripts encoding proteins that regulate a broad range of biological activities. They have a capacity to effect tissue repair via multiple mechanisms including promoting the survival and proliferation of endogenous cells, inhibiting inflammatory and immune responses (Phinney & Prockop 2007). MSCs release factors that not only enhance the regeneration of non-haemopoietic tissues, but also yield proangiogenic factors essential for HSCs/EPCs survival, proliferation and differentiation (Kinnaird et al. 2004; Honczarenko et al. 2006), thus having a role in blood vessel formation as well. Wounds receiving implanted MSCs also show enhanced angiogenesis (Chen et al. 2008), which further indicates their importance during wound repair.
Recovery of blood flow at an injury site is considered to be a prerequisite for tissue reconstitution. Vasculogenesis, the de novo process by which progenitor stem cells differentiate and give rise to a replacement vascular network, was earlier believed to take place only during embryonic development. Bone-marrow-derived endothelial progenitors have, however, been identified in the peripheral blood and shown to participate in post-natal new blood vessel formation (Velazquez 2007). The granulation tissue of the wound is, thus, vascularized by the processes of angiogenesis or vasculogenesis, or both.
Here, we noted that the HA-coated sponges had an increased number of cells positive for c-kit and CD34, both known to be markers for endothelial lineage cells, and more stro-1-positive cells indicating more abundant MSCs. The more vascularized granulation tissue in HA-coated sponges (Tommila et al. 2008) might be a result of more abundant homing of bone-marrow-derived stem cells to HA-coated cellulose.
To conclude, HA-coated cellulose sponges used to induce and harvest granulation tissue seem to have cell-guiding features that facilitate homing of bone-marrow-derived stem and progenitor cells more efficiently than untreated cellulose. This property of HA-coated cellulose together with its capacity to promote proliferation of richly vascularized connective tissue (Tommila et al. 2008) might have potential clinical application in the treatment of poorly healing wounds.
The study design was approved by the Animals Ethical Committee of the State Provincial Office of Southern Finland.
Maria Ström is acknowledged for skilful technical assistance. This study was financially supported by the Magnus Ehrnrooth Foundation (A.J.), and the Finnish National Technology Agency (TEKES, no. 40197/06).