Rhizoma Curculiginis and Rhizoma Drynariae were purchased in a local Chinese medicine store and were identified morphologically, histologically and chemically according to standard Chinese herbal identification procedures 2021. Initially, the morphology and histology of the herbs were compared with standard photographs obtained from the School of Traditional Chinese Medicine, the University of Hong Kong. The actual identification procedures were performed by the authors in the Hard Tissue Laboratory, the University of Hong Kong. Thin layer chromatography was used to separate the components of Rhizoma Curculiginis ethanolic extract. Potassium ferrocyanide (2%) and ferric chloride were added to the extract, which produced bluish spots (compared with the standards), an indication of the authenticity of the herb 21. Methanolic extract of Rhizoma Drynariae was used for thin layer chromatography. The extract was separated by benzene-methanol-butanone (at a ratio of 3:1:1) and tested with ferric chloride-ethanol solution (1%). Brownish colored extract was compared with naringin standard 20. A sample of each herb was stored in the Hard Tissue Laboratory. Rhizoma Curculiginis and Rhizoma Drynariae extracts were prepared according to the protocol for commercial production of injection preparation of traditional Chinese medicine in China 22. For every 4 g of Rhizoma Curculiginis or Rhizoma Drynariae powders, 40 ml of distilled water was added and the mixtures were boiled with stirring on a hot plate for 4 hours. Distilled water was added occasionally to prevent the mixtures from drying. The final volume of the mixtures was made up to 4 ml by adding distilled water. The mixtures were cooled to room temperature and then centrifuged. The supernatants were collected and filtered with a 0.22 μm sterile syringe filter into a sterile glass bottle. The extracts contained 1 g/ml of Rhizoma Curculiginis or Rhizoma Drynariae. This method of extraction is widely used for obtaining water soluble fractions in Chinese medicinal herbs 23.

All animals were obtained from the Laboratory Animal Unit at the University of Hong Kong where they were kept under standard conditions. The temperature was kept between 22°C and 24°C. The light cycle was from 8 am to 8 pm daily. Each animal was kept individually in a cage and fed with standard diet in the Laboratory Animal Unit. The animal handling and experimental protocol was approved by the Committee for the Use of Living Animals in Teaching and Research, the University of Hong Kong.

The mice were bred by the Laboratory Animal Unit, the University of Hong Kong. The doses of both Rhizoma Curculiginis and Rhizoma Drynariae for human were 0.2 g/kg/day as suggested in the Chinese Pharmacopeia 2. The doses of the herbs for mice were estimated in a pilot study. Thirty 8-week-old male BALB/c mice were divided into three groups as follows:

1. Control group (C_mice): Ten BALB/c mice fed with normal diet and distilled water.

2. Rhizoma Curculiginis group (XM_mice): Ten BALB/c mice fed with normal diet and distilled water mixed with Rhizoma Curculiginis extract (0.5 g Rhizoma Curculiginis in 250 ml of water).

3. Rhizoma Drynariae group (G_mice): Ten BALB/c mice fed with normal diet and distilled water mixed with Rhizoma Drynariae extract (0.5 g Rhizoma Drynariae in 250 ml of water).

The mice were kept individually (one animal per cage) for five weeks before sacrificed. The drinking solutions, which did not contain any suspending herbal particulates, were freshly prepared every day.

The bone samples (proximal tibia) were carefully dissected after sacrifice of the mice. The sample was then fixed with buffered saline and placed onto a sample holder. The proximal tibia is a long bone which can be easily inserted in the micro-CT chamber, allowing easy standardization in specimen positioning. These procedures were originated from other similar micro-CT studies 18. The investigator was blinded to the treatment of each mouse. The morphometric parameters were determined automatically by computer without human interference. The bone samples were scanned through 360° by a compact fan-beam-type tomography instrument μCT20 (Scanco Medical AG, Bassersdorf, Switzerland). Samples were placed in an air-tight cylindrical sample holder filled with formaldehyde to preserve the sample for the duration of the measurement. The sample holder was marked with an axial alignment line on the outside of the tube to allow consistent positioning of the specimens within the holder. By connecting the alignment notch to the sample holder with the μCT20 turnable, precise positioning of the bone within seconds was achievable. A typical analysis consisted of a scout view to ensure accurate and consistent positioning of slides, selection of the examination volume, automatic positioning, measurement, offline reconstruction and evaluation. Scout (enlarged) views were taken to ensure the precision of finding the anatomic location in each bone, a sample size estimated from similar studies was selected to minimize the effect of the location variation.

Modified protocols from Zhang et al. 24 and Ishimi et al. 18 were followed: twenty micro-CT slices/sections (0.25 mm apart) were taken to cover the proximal end of the left tibia. The most proximal slice was 1.5 mm away from the proximal end to avoid possible morphological variation in the proximal head (Figure 1). Quantitative morphometry of the bone structure was performed with the μCT20 computer system. The fibula was excluded from measurement. The micro-CT reconstruction parameters were as follows: sigma: 1.2; support: 2; threshold: 140; increment/scan thickness: 0.25 mm; resolution: high (1024 × 1024 pixels).

Data were analyzed using statistical analysis software Graphpad Instat (v.2.04a). The arithmetic mean and standard deviation (SD) were calculated for each group. The means (XM_mice and C_mice; G_mice and C_mice) were compared by the Welch's unpaired t test which does not assume equal variances, with P < 0.05 chosen as the critical level of statistical significance.

The methodology and animal model used in the current study have been described previously 11. Six 10 × 5 mm² full-thickness bone defects were created in the parietal bones of three inbred New Zealand white rabbits. The rabbits were five months old (adult stage) and weighed between 3.5 kg and 4.0 kg. The animal handling and experimental protocol was approved by the Committee on the Use of Live Animals in Teaching and Research, the University of Hong Kong. In the experiment, two defects in the first animal were grafted with collagen matrix with Rhizoma Curculiginis extract; two defects in the second animal were grafted with collagen matrix with Rhizoma Drynarie extract; two defects in the third (control) animal were grafted with collagen matrix alone. The animals were pre-medicated one hour before surgery with oxytetracycline hydrochloride (200 mg/ml, 30 mg/kg body weight) and buprenorphine hydrochloride (0.3 ml/kg body weight), supplemented with diazepam (5 mg/ml, 1 mg/kg body weight). In order to maintain the level of neuroleptanalgesia, increments of Hypnorm (0.1 ml/kg) were given at 30-minute intervals during the operation.

The surgical procedure consisted of the creation of two 10 × 5 mm full-thickness (approximately 2 mm) cranial defects, devoid of periosteum, using templates, in the parietal bones. The defects were produced using round stainless steel burs (1 mm in diameter) on a low speed dental drill. Outlines of the defects were made initially by making holes of full thickness the parietal bone using a stainless steel wire template bent to the required size of the defect. The holes were joined to complete the process. During the cutting of the bones, copious amount of sterile saline was used for irrigation and to minimize thermal damage to the tissues. In the first experimental animal, the defects were filled with purified absorbable fibrillar collagen matrix (Collagen Matrix, NJ, USA) with 0.2 ml 1 g/ml Rhizoma Curculiginis extract. In the second experimental animal, the defects were filled with the same collagen matrix with 0.2 ml 1 g/ml Rhizoma Drynarie extract. The grafts were prepared 15 minutes before grafting. In the control animal, the defects were grafted with 0.02 g of the same collagen matrix mixed with 0.2 ml water for injection.

All wounds were closed with interrupted 3/0 black silk sutures. No attempt was made to approximate the periosteum to prevent the barrier effect. Postoperatively, the rabbits were given oxytetracycline hydrochloride daily for ten days and buprenorphine hydrochloride for two weeks. The animals were monitored under a standardized protocol during postoperative period with the supervision of the veterinary surgeon for any unhealthy signs or side effects. Medication for the animals during this period was as follows: 30 mg/kg of oxytetracycline hydrochloride intramuscular injection every 30 minutes; 50 μg/kg of Temgesic subcutaneous injection daily (for two weeks); 60 ml of saline and 10 ml Dextran 40 (10% dextrose solution in 0.9% saline solution) subcutaneous injection daily until appetite recovered; 10 ml of Dextran 40 in 350 ml drinking water daily until appetite recovered; vitamin B1, B6 and B12.

Two weeks after surgery, the animals were sacrificed with sodium pentobarbitone. Immediately after death, defects and surrounding tissues were removed for histological preparation.

Tissues were fixed in 10% buffered saline solution, demineralized with K's Decal Fluid (sodium formate/formic acid) and double embedded in celloidin-paraffin wax. Serial, 5-μm-thick sections of the whole defect were cut perpendicular to the long axis. The slides were stained with Periodic acid-Schiff stain which allowed easy identification of new bone formation.

For the investigation of the ultra-structure of new bone formation, some tissues of the animal grafted with Rhizoma Drynarie extract was fixed with Karnovsky solution, decalcified in K's Decal Fluid and double embedded in celloidin-paraffin wax. The specimens were cut into 3-μm-thick serial sections, perpendicular to the long axis, and stained with hematoxylin and eosin. Histological sections were used to locate the area of the interface between the graft and the host bone on the tissue blocks. The areas of interest were then fixed, post-fixed in osmium tetroxide (OsO₄) and trimmed into 1 mm blocks and embedded in resin. Semi-thin sections from resin block were sectioned and stained with toluidine blue for further orientation and reduction of the block face. Ultra-thin sections (90 nm thickness) were cut with a diamond knife to and mounted on metal grids (200 meshes). Sections were then stained with uranyl acetate and lead citrate, and examined under a transmission electron microscope (EM208S, Phillips Electron Optics BV, Netherlands).

All animals remained in excellent health throughout the course of the experiment. There was no evidence of side effects in any of the animals. The mean (SD) weight of the mice was 28.46 (1.03) g and there was no significant difference in weight between the experimental and control groups before and after the experiment. The mean (SD) daily water consumption of each mouse was 6.6 ml. There was no significant difference in daily water consumption between the XM_mice: 6.7 (1.1) ml, G_mice: 6.5 (1.1) ml and C_mice groups: 6.6 (1.3) ml.

Micro-CT scanning was performed on the proximal tibia of the XM_mice, G_mice and C_mice groups. 2D-histomorphometry was calculated from the micro-CT system. Compared with the C_mice control group, trabecular number increased 10.00%; trabecular thickness and trabecular separation decreased 7.01% and 9.66% respectively in the XM_mice group. The low trabecular separation is consistent with the high trabecular number, i.e. finer trabecular pattern with more and thinner trabeculae. In the XM_mice group, the low trabecular separation is also consistent with increased bone surface to tissue volume and bone surface to bone volume ratios which went up 10.03% and 6.68% respectively. In the G_mice group, compared with the C_mice control group, trabecular number increased 10.00%; trabecular thickness and trabecular separation decreased 3.78% and 9.66% respectively. Rhizoma Drynariae extract caused a lower trabecular thickness, which was consistent with the overall increase in the bone volume/tissue volume ratio (up 6.45%), an indicator of bone density (Table 1).

For method error analysis, readings of ten sections recorded at two separate micro-CT scanning sessions by the same examiner were randomly drawn and the readings were then compared. The method error of the image analysis (∑d22n MathType@MTEF@5@5@+=feaagaart1ev2aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGacaGaaiaabeqaaeqabiWaaaGcbaqcfa4aaOaaaeaadaWcaaqaamaaqaeabaGaemizaq2aaWbaaeqabaGaeGOmaidaaaqabeqacqGHris5aaqaaiabikdaYiabd6gaUbaaaeqaaaaa@3339@) is 0.003194 mm², where d is the difference between the two readings of a pair and n is the number of double readings. The error was insignificant compared with the results. The duplicate intra-observer registrations of the 10 randomly drawn sections were also insignificant (P = 0.1679).

All animals remained in excellent health throughout the course of the experiment and recovered rapidly after operation. There was no evidence of side effects or infection in any of the animals.

In the rabbits grafted with Rhizoma Curculiginis or Rhizoma Drynariae extracts in collagen matrix, new bone was formed at the host bone – graft interface tended to grow across the defect (Figures 2 and 3). Integration of the Rhizoma Curculiginis or Rhizoma Drynariae extracts and collagen with the recipient bed was characterized by the appearance of new bone. No cartilage was found. At higher magnification, new bone could be seen growing around a collagen matrix fragment and tended to grow towards and amalgamate with the collagen matrix. The presence of bone cells indicated that new bone was formed.

In the control rabbit, little new bone was formed at the host bone/graft interface. Some collagen fibers were present at the center of the defects (Figure 4). In all rabbits, the defects were healed, with fibrous tissue bridging across the defects.

The area of interest for ultra-structural study was taken near the newly formed bone and its surrounding connective tissue. At the mineralization front (F), osteoblasts (OB) with abundant rough ER and newly entrapped osteocytes (OC) could be seen (Figure 5).