CORRELATION BETWEEN MICRO-CT SECTIONS AND HISTOLOGICAL SECTIONS OF MOUSE SKULL DEFECTS IMPLANTED WITH ENGINEERED CARTILAGE

Abstract

One advantage of using cartilage to replace/repair bone is that the implant disappears as bone is formed by endochondral ossification. Previously, we showed that cartilage spheroids, grown in a rotating bioreactor (Synthecon, Inc.) and implanted into a 2 mm skull defect, contributed to healing of the defect. Skulls with or without implants were subjected to microCT scans. Mineralized regions from microCT sections correlated with regions of bone in histological sections of the defect region of demineralized skulls. Recently, sections from microCT scans of live mice were compared to histological sections from the same mice. The area of the defect staining for bone in histological sections of demineralized skulls was the same region shown as mineralized in microCT sections. Defects without implants were not healed. This study demonstrates that microCT scans are an important corollary to histological studies evaluating the use of implants in healing of bony defects.

INTRODUCTION

For almost 20 years, my lab has used NASA’s rotating bioreactors to grow bone-forming cartilage (Duke et al., 1993). Initially, we focused on the bioreactor’s ability to simulate microgravity but soon turned to tissue engineering studies that exploited the bioreactor’s capabilities as a 3D tissue culture apparatus (Duke et al., 1996).

Tissue engineering is a possible solution to the problem of procuring enough bone for craniofacial repair. In most tissue engineering studies, bone cells (osteoblasts) are cultured on a scaffold with or without the addition of growth factors, and then implanted; however a few labs including our own have been using cartilage systems for bone repair (Oliveira et al., 2009A, 2009B; Jukes et al., 2008). The rationale behind these studies is that most of the bony skeleton forms embryonically through the process of endochondral ossification in which a cartilaginous template mineralizes, then vascularizes and subsequently forms bone (Duke et al., 1993; Duke et al., 1996). Fracture healing also includes a cartilaginous stage (Urist, et al., 1967).

In one study, 14–day chick sternal chondrocytes were cultured in chitosan sponges, treated with retinoic acid to induce cartilage maturation, then implanted subcutaneously in nude mice. Vascularization and new bone deposition in and around the cartilage/chitosan template was induced (Olivera, et al., 2009A, 2009B). Another group placed cells from an embryonic mouse stem cell line on a carrier, differentiated them into cartilage, and implanted them into a defect in nude rats. Extensive bone formation was observed (Jukes et al., 2008).

Our lab has used tissue engineered cartilage to replace or repair bone, with the major difference from other studies being that we do not use a scaffold (Duke, et al., 1993). Instead, to produce cartilage nodules to be grown in the bioreactor, we use an old technique developed to study cell adhesiveness--aggregation of single limb bud mesenchymal cells and their subsequent differentiation into bone-forming cartilage (Moscona, 1961; Duke and Elmer, 1977). Cartilage nodules engineered this way were implanted subcutaneously and shown to form bone, and later implanted into a defect created in the skulls of mice (Montufar-Solis, et al., 2004).

To assess mineralization, skulls with or without implants were imaged using microCT scanning (Figures 1A; 2A and B). (Superior color versions of these pictures are available at http://www.db.uth.tmc.edu/orthodont/Duke%20NASA/engineered_cartilage.htm).

Figure 1.

A. MicroCT scan of control skull #10 showing plane of section for S1. S2 is another microCT section of the same defect. Arrows indicate edge of defect. B. Histological section of a control skull. P = periosteum. Bar = 2 mm. Arrows indicate edge of defect.

Figure 2.

Skull with implant. A. MicroCT scan showing healed defect. B. MicroCT scan indicating section shown in C. D. Histological section. p = periosteum; b = bone. Arrows indicate edge of defect.

The skull was then demineralized, and the defect region, with or without an implant, was sectioned and stained with an Alcian Blue-Tartrazine stain which differentially stains cartilage and bone. The microCT scan showed that mineralization was occurring. This was confirmed by histology showing that the implant consisted of a mixture of cartilage and bone, and that the amount of cartilage diminished over time as the amount of bone increased. Vascularization of the tissue occurred concomitantly with the formation of bone.

The cartilage spheroids, implanted into a skull defect, contributed to healing of the defect (Figure 2). In skulls which had the defect but no implant, some sections exhibited only a thickened periosteum on either side of the defect or a layer of bone that was thinner than the surrounding region, with a thickened periosteum (Figure 1B). In contrast, the defect in implanted skulls was filled with a mass of bone (Figure 2D). Location of bone in histological sections correlated with location of mineral in the microCT sections (Duke et al., 2008).

In the current report, microCT sections of live mice with or without implants were compared to histological sections obtained after euthanasia of the mice at 2 and 4 weeks.

MATERIALS and METHODS

All animals used in these experiments were handled in compliance with the Animal Welfare Act and NIH Policy and Guide, and all procedures were approved by the UTHSC-Houston Animal Welfare Committee. Imaging procedures were approved by the MD Anderson Cancer Center Animal Welfare Committee as well.

Cartilage for implants

Pre-experimental procedures included mating of C57BL/6NH black mice (Harlan-Sprague-Dawley, Inc.) and weighing the females to confirm pregnancy. Pregnant females were euthanized by cervical dislocation; E13 or E13.5 embryos were extracted, placed in Tyrodes, and fore and hind limb buds removed, minced, trypsinized, mechanically dissociated in culture medium and filtered through Nitex (20 µm) or a 70 µm Cell Strainer (BD Falcon) to form a single cell suspension. The medium was BGJb-Fitton-Jackson modification (GIBCO) supplemented with 10% fetal bovine serum (Fischer), 150 µg/ml ascorbic acid, 1% Penicillin-Streptomycin solution (Sigma), and 0.25 µg/ml Fungizone (GIBCO). Two ml of cell suspension (3.5 × 10⁶ cells/ml) were placed per compartment of a polystyrene divided Petri dish (Falcon) on a rotating platform (American Rotator V R4140) inside a humidified incubator (37°C and 5% CO₂) and aggregated overnight. The next day, the aggregates were inserted into 50 ml commercially available bioreactor vessels (Synthecon, Inc.) and cultured for three weeks inside the same humidified incubator (5% CO₂). Bioreactor speed was set initially at 23.3 rpm. One half of the culture medium was changed every two days, and speed of the bioreactor was increased as necessary to maintain cartilage nodules in suspension. After 3 weeks, cartilage nodules were harvested and used for histology or implantation.

Histology

Cartilage nodules for histological studies were fixed in 10% buffered neutral formalin, embedded in paraffin, sectioned and stained either with Toluidine blue for assessment of cartilage differentiation and matrix production or with von Kossa or alizarin red for assessment of mineralization (Lillie and Fulmer, 1976). Toluidine blue is a metachromatic dye which undergoes a color change to pink or purple in the presence of the sulfated glycosaminoclycans in cartilage matrix Von Kossa staining is used to indicate mineralized tissue, because the stain reacts with phosphates present in the mineral. Alizatin red is used to detect calcium.

Implants

Mice used for implants were the same strain that provided cells for the cartilage spheroids cultured in the bioreactor---C57BL/6NH. Eight-week-old males were initially anesthetized with 4% isoflurane; then, for the duration of the procedure, maintained on 1.5% isoflurane. Heads of anesthesized mice were shaved at the incision site and cleaned with a chlorhexidine solution. A subcutaneus injection with the local anesthetic Bupivacaine was given before a 1 cm incision was made above the left parietal bone and the skin separated. Under saline irrigation, a 2 mm circular defect was drilled in the superior portion of the left parietal bone using a Trephine bur attached to a Micro Drill handpiece (11,000 rpm, Fine Scientific Tools, Inc.). Great care was taken not to damage the underlying dura. The circular piece of bone was removed and a cartilage nodule was implanted in the defects of six mice. Six mice were operated but did not receive implants. The incision was closed with 8-0 Nylon swaged microsurgical sutures (S&T). All mice were treated immediately post-surgery and on each of the 3 following days with Tritop (Pharmacia and Upjohn) ointment, which has antiseptic and anesthetic effects.

Analysis

Histology of implants

One mouse of each treatment was sacrificed after 2 weeks and 4 weeks of implantation. Skulls were removed and fixed, and the area containing the hole and/or implant was removed and decalcified in EDTA, then processed for histology. Bone formation was assessed by Alcian blue-Nuclear Fast Red-Tartrazine, which stains nuclei red, bone matrix yellow, cartilage blue, bone canaliculi blue-green, and cytoplasm light pink (Sams and Davies, 1967).

Micro-CT Scans

To assess healing over time, the mice received in vivo microCT scans at weeks 2 and 4 in order to visualize the appearance of new calcified tissue. The scans took place at the Small Animal Imaging Facility (SAIF) in the Department of Imaging Physics at M.D. Anderson Cancer Center, and all animal procedures were approved by and performed under the guidelines of The University of Texas M. D. Anderson Cancer Center Institutional Animal Care and Use Committee.

The system used was a model Locus RS-9 tabletop cone-beam micro CT scanner (General Electric Medical Systems, London, Ontario). The scanner gantry rotates in a step-and-shoot manner around the object bed, imaging the animal in a single pass. The gantry features a tungsten anode x-ray source fixed opposite a cesium iodide scintillator and a charge-coupled camera. Each mouse was positioned prone on a half-pipe acrylic sled with a nose cone attached by plastic tubing to a vaporizer set to deliver 2% isoflurane (Halocarbon Laboratories, River Edge, NJ) with an oxygen flow rate of 2 liters/min. For this free-breathing animal protocol, 360 projection images were taken at one-degree increments, with each projection consisting of a 500-msec X-ray exposure with the X-Ray source operated at 80 kVp and 470 µA.

After the raw acquired images were normalized against maximum and minimum x-ray exposure frames, a backprojection reconstruction method was used (i.e., bright and dark fields), correcting for any bad pixels in the detector and scaling the raw CT numbers into Hounsfield units (HU). With proper calibration, air has a value of −1000 HU, water has a value of 0, and bone (pure hydroxyapatite) has a value of approximately 2,700 HU.

The GE MicroView CT program was used to generate 3D renderings of the mouse skulls based on a minimum threshold signal (600 HU), with any signal in the scan at or above that threshold displayed. 600 HU was the user-determined lowest value that corresponded to calcified bone without incorporating non-mineralized soft tissue as well.

Correlations of sections

Histological sections from near the center of the defect region stained as described above were selected for comparison with microCT sections. The microCT sections were chosen by using the GE Microview CT program software to move through sections of the scan until the center of the defect was reached. An image of the section was then captured and compared with the histological sections.

RESULTS

Fixed cartilage nodules (diameter 1.5–2mm) are shown in Figure 3. Note their irregular appearance, which results from the combining of several smaller nodules. Within the nodules, chondrocytes differentiated and hypertrophied as shown in Figure 4. Abundant metachromatic cartilage matrix is present as are chondrocytes in various stages of differentiation. Mineralization of the nodules is shown by the von Kossa staining in Figure 5.

Figure 3.

Nodules (about 1.5mm–2mm) fixed with formalin prior to histological analysis.

Figure 4.

Toluidine blue stain demonstrates cartilage matrix in nodule sections. The nodule contains primarily hypertrophic cells (white arrow) and is surrounded by a perichondrium (black arrow). Bar = 34.5 µm.

Figure 5.

Nodule stained with Von Kossa showing mineralization (areas of black deposition). Bar = 36.3 µm.

Demineralized defect regions were sectioned for histological studies and stained with an Alcian Blue-Tartrazine stain which differentially stains cartilage and bone. These sections are shown in Figures 6–9 along with their corresponding microCT scans. In each of these specimens, when the microCT section is compared with the histological section, the regions of bone in the histological section, which stain yellow and appear lighter in the figure, are shown to be mineralized in the microCT scan.

Figure 6.

Skull with no implant, 2 weeks post-surgery. A. Plane of section of µCT scan. B. µCT saggital section of defect region. C. Histological section of defect region stained with Alcian blue-Nuclear fast red –Tartarazine stain. Arrow = 310 µm.

Figure 9.

Skull with implant, 4 weeks post- surgery. A. Plane of section of µCT scan. B. µCT saggital section of defect region. C. Histological section of control skull. B = bone. M = marrow. Arrow = 342 µm.

In controls, which had a defect with no implant, the defect was still present at both 2 and 4 weeks (Figures 6 and 7). In the histological sections, no tissue was seen within the defect itself (Figure 6C, 7C) and the microCT scans (Figures 6B and 7B) contain no mineralized tissue in the gap.

Figure 7.

Skull without implant, 4 weeks post surgery. A. Plane of section of µCT scan. B. µCT saggital section of defect region. C. Histological section of skull without implant. Arrow = 300 µm.

In contrast, the defect in implanted skulls at 2 weeks was filled with a mixture of cartilage and bone (Figure 8C), and the microCT scan showed the presence of mineralized tissue within the gap (Figure 8B). The blurriness of the microCT image is due to the mouse moving during the scan. Still, regions staining yellow with tartrazine followed the same pattern as the mineralized tissue in the uCT scans.

Figure 8.

Skull with implant, 2 weeks post surgery. A. Plane of section of µCT scan. B. µCT saggital section of defect region. C. Histological section of skull with implant. Arrow = 330 um. Arrowheads indicate edges of defect.

The implant placed in the mouse sacrificed at 4 weeks did not fit neatly into the defect, and not only filled the defect, but extended below the plane of the skull into the brain (Figure 9). In this implant, at 4 weeks, there was no cartilage remaining, and the implant consisted primarily of bone and marrow. The microCT scan shows the protrusion of the implant into the brain, and again the presence of mineralized tissue in areas shown in the histological sections to be bone.

DISCUSSION

The primary purpose of the overall study was to show that cartilage can heal defects created in a membranous bone in the skull of a mouse. As before, the healing was shown to progress through the stages of endochondral ossification, which occurs during embryonic development and fracture repair and forms bone through replacement of a cartilaginous template. During this process, chondrocytes proliferate, hypertrophy, and secrete alkaline phosphatase, responsible for calcification (Adams et al., 2007). As calcified matrix surrounds the hypertrophied cells, the cells undergo controlled disintegration via apoptosis. Apoptosis in growth plates has been shown to generate conditions favorable to calcification and vascularization (Duke et al., 2003). Finally, the developing structure forms a periosteum, a haven for blood vessels and undifferentiated cells, and bone continues to form replacing the cartilage template. This process is the same as that described by Urist and others (1967) and by Reddi and coauthors (1987) for the formation, through an endochondral process, of ectopic bone due to implantation of demineralized bone matrix.

Because we have concentrated on the bone healing capabilities of our nodules, their molecular aspects have not yet been addressed, but histologically, cells in our nodules appear to undergo a typical chondrogenesis and subsequent chondrocyte differentiation, as seen in Figures 3 and 4 above. In our in vitro system, the endochondral process cannot be completed, since there is no source to contribute blood vessels. In vivo, however, this occurs as vascularization of the mineralized implant is apparent in the decalcified skull prior to sectioning (data not shown), and can be seen in sections of the skull containing the implant. (See Figure 8C.) Eventually, all of the cartilage becomes replaced with bone containing marrow spaces (see Figures 8C and 9C). To definitively determine whether this bone forms via induction from the host, a series of experiments using green fluorescent protein (GFP) mice are planned.

The issue of vessel formation is particularly important as it is a limiting factor in integration of implanted bone or engineered constructs. Using cartilage removes this barrier as mineralized cartilage induces angiogenesis (Montufar-Solis, et al., 2004).

In contrast to our previous study, which showed some bridging of the gap in the control skulls (see Figure 1), no mineralized tissue was seen in the microCT scans of the controls or in their histological sections. This difference is likely due to the use in the previous experiments of Millipore filters to hold the implant in place. Without the filter being present, the periosteum was not able to bridge the gap as readily. Recently completed scans of animals one year post-surgery showed that the defect was still present.

Our studies have shown the ability of bone-forming cartilage to heal a defect in a membranous bone, but the problem remains of how to shape the implant more effectively so that it will more closely follow the shape of the skull, which in the case of the cranium, as used in our studies, means being fairly flat, or possibly just slightly curved, certainly not spherical.

We are also exploring the use of bone marrow derived mesenchymal stem cells as a source of cartilage for implants. This cell source is especially attractive in the case of craniofacial defects in children, e.g. cleft palate, or Treacher-Collins syndrome, which require multiple surgeries throughout their lives (Duke et al., 2009). These are exciting applications of a space technology to problems on Earth.

CONCLUSION

In order to carry out studies on the timing of healing in mice implanted with engineered cartilage, we performed microCT scans of living mice, which were sacrificed at later dates, and the implanted skull region removed and prepared for histology. In histological sections of demineralized defect regions, the area of the defect staining yellow with tartrazine, indicating the presence of bone, was the same region shown to be mineralized in microCT sections. Defects without implants were shown in microCT sections, as in histological sections, to not be healed. This study and the previous one demonstrate that microCT scans are an important corollary to histological studies evaluating the use of implants in healing of bony defects.