Effect of antiresorptive and anabolic bone therapy on development of osteoarthritis in a posttraumatic rat model of OA

Animals and management

Four-month-old male Lewis rats (Charles River Laboratories, Portage, MI, USA) weighing
approximately 350 g at the beginning of the experiments were used in this study. All
in vivo procedures were approved by the Institutional Animal Care and Use Committee
(IACUC) at Pfizer (Andover, MA, USA) and were performed in accordance with the Guide
for the Care and Use of Laboratory Animals and with the US National Institutes of
Health (NIH) Publication No. 85-23, revised in 1996 35]. All efforts were made to minimize animal pain and distress and to minimize the number
of animals used. The rats were pair housed in ventilated Innovive cages in a temperature-
and humidity-controlled room on a regular 12-hour light/dark cycle. Irradiated LabDietâ„¢
5053 (Purina, Richmond, IN, USA) and water were provided ad libitum. The rats were
acclimated for 1 week prior to use in the study. A total of 48 rats were used for
the 10-week study, with 12 rats per group. The rats were randomized to four study
groups based on their body weights on the day before surgery. A group of 12 rats received
sham surgery, and another group of 36 rats underwent MM surgery. The four study groups
were as follows: sham control group (Sham), MM vehicle-treated group (MM?+?veh), MM?+?zoledronate
group (MM?+?Zol) and MM?+?PTH group (MM?+?PTH).

Surgery

The rats were induced and maintained under anesthesia using isoflurane. One dose of
carprofen (Pfizer Animal Health, New York, NY, USA) and sustained-release buprenorphine
(Zoopharm, Windsor, CO, USA) were administered prior to surgery for analgesic coverage.
The right knee was shaved, aseptically prepared, and draped for surgery. In the sham
group, a surgical approach to the medial collateral ligament (MCL) on the right hind
limb was completed by only cutting the skin while the MCL was left intact. The surgical
incision was closed in two layers using absorbable sutures. In the surgery groups,
medial meniscectomy (MM) was performed by fully transecting both the MCL and the medial
meniscus of the right hind limb, followed by closure in two layers using absorbable
sutures 36]. All of the rats were euthanized at 10 weeks post surgery.

Dosing and bone labeling

The rats in the Sham and MM?+?veh control groups (groups 1 and 2) received vehicle
(sterile water used for injections) at 1 mL/kg subcutaneously (sc) 5 days/week, starting
on the day of surgery. The rats in group 3 received zoledronic acid (Zol; Sargent
Pharmaceuticals 25021-801, Schaumburg, IL, USA) at 100 ?g/kg, sc, twice/week 18], and the rats in group 4 received human PTH (hPTH 1-34; Sigma-Aldrich P3796, St.
Louis, MO, USA) at 40 ?g/kg, sc, five times/week 25], starting on the day of surgery. To label actively mineralized bone surfaces, all
of the rats in the study received calcein (Sigma-Aldrich Cat. No. C-0875, St. Louis,
MO, USA) at 10 mg/kg (3.3 mL/kg) 13 days before necropsy and alizarin (Sigma-Aldrich
Cat. No. A-5533, St. Louis, MO, USA) at 30 mg/kg (3.0 mL/kg) 3 days before necropsy.

Body weight, sample collection and serum analyses

Body weight was recorded twice weekly throughout the study. At the end of the 10-week
study the entire right hind limb was carefully harvested, skinned, and cleaned of
the soft tissue, with care taken not to disrupt the knee joint. The limbs were wrapped
in saline-soaked gauze and frozen at -20 °C for ex vivo imaging and histological analyses
of the tibial articular cartilage and bone. Blood was collected by jugular venipuncture
under isoflurane anesthesia. Serum was analyzed for biomarkers of bone formation and
bone and cartilage degradation at 5 and 10 weeks post surgery. Osteocalcin was assessed
by rat EIA kit (Cat. No. BT-490, Biomedical Technologies, Stoughton, MA, USA) and
procollagen type 1 N-terminal propeptide (P1NP) was quantified in serum samples by
the liquid chromatography/mass spectrometry (LC/MS) method 37]. Cross-linked type I collagen C-terminal telopeptide (CTX-I) was analyzed by rat
LAPS™ Assay (Cat. No. AC-06 F1) and C-terminal telopeptide of type II collagen (CTX-II)
by CartiLapsâ„¢ Assay (Cat. No. AC08F1) all produced by Immunodiagnostics Systems, Scottsdale,
AZ, USA.

At the end of the study serum chemistry analysis included alanine aminotransferase
(ALT, U/L), glutamic-oxaloacetic transaminase (AST, U/L), alkaline phosphatase (ALP,
U/L), gamma-glutamyl transpeptidase (GGT, U/L), total bilirubin (TBIL, mg/dL), cholesterol
(CHOL, mg/dL), glucose (GLUC, mg/dL), total protein (TP, g/dL), albumin (ALB, g/dL),
globulin (BLOB, g/dL), albumin/globulin ratio (AG, ratio), blood urea nitrogen (BUN,
mg/dL), creatinine (CREA, mg/dL), phosphorus (PHOS, mg/dL), calcium (Ca, mg/dL), sodium
(Na, mmol/L), potassium (K, mmol/L) and chloride (Cl, mmol/L).

Dynamic weight bearing

Dynamic weight-bearing (DWB) measurements were obtained before surgery, at week 5
and before euthanasia to assess the effects of surgery on the weight-bearing capacity
of the hind and front legs. The DWB system (model BIO-SWB-R, Bioseb, Boulogne, France)
is a noninvasive method used to obtain the weight and surface area of all four feet
in a freely moving animal 38]. The system consists of a Plexiglas enclosure (22?×?22?×?30 cm) with a floor sensor
consisting of an array of 44?×?44 sensors. A camera is affixed to the side of the
enclosure to align the sensor directionally during the analysis phase. Raw pressure
and live video recordings are transmitted to a tablet PC via a USB interface at a
sampling frequency of 10 Hz. The data were analyzed using the DWB software, version
1.3. Zone parameters were set for the analysis as follows: ? 4 g for one sensor or
a minimum of three adjacent sensors???2 g (to be considered a valid zone). For each
time segment that was stable for more than 1 second, zones that met the above criteria
were validated and assigned as either right or left and front or rear. Mean values
for the weight and area of each zone were calculated over the entire testing period
based on the length of time for each validated segment. For each testing period, the
animals were placed into the chamber and allowed 20–30 seconds to explore prior to
the data collection time of 3 minutes. The operator manually “validated” each test
period by ensuring that each paw print corresponded to the appropriate paw, using
the video footage as a reference. The following parameters were measured for each
leg separately and for both the front and hind legs combined: body weight (g), weight
put on the limb (g), percentage of weight placed on the limb and surface area of each
paw (mm
²
).

Radiology

All of the bone samples were X-rayed with the Faxitron Model MX20 specimen scanner
(Faxitron Bioptics LLC., Tucson, AZ, USA) using the settings recommended by the manufacturer;
exposure time 12–18 seconds at 31–35 kV. All of the samples were imaged at 3× magnification
and were positioned horizontally with the center of the beam at the knee joint. Both
frontal and lateral views of each sample were obtained. Radiographic images were used
to assess the gross anatomy of the region of interest to be evaluated by ?CT and to
inspect the bone samples for the presence of fractures or other bone abnormalities.

?CT and EPIC ?CT measurements

?CT was conducted on the right knee joint, utilizing a MicroCT 100® computed tomography
system (Scanco Medical, Bassersdorf, Switzerland) to obtain a scout three-dimensional
image of the knee and ensure that the samples were reproducibly scanned and analyzed
exactly at the same region of interest (ROI) in each specimen and that the size of
the ROI that we selected allowed for meaningful analysis of bone structures at the
proximal tibial epiphysis and metaphysis.

Following imaging of the entire knee, the femur and tibia were carefully separated
to ensure that the articular cartilage and meniscus of the joint were not disrupted.
The tibia was then cut above the tibiofibular junction and the proximal tibia was
placed in a plastic custom-made positioning device to ensure consistent scanning.
Precontrast scans of all the tibias were obtained using the MicroCT 100® with the following parameters: 800 slices, a 10-?m resolution, a total scanned area
of 8.0 mm
²
, and source energy of 70 kVp, 115 ?A at 8 W to capture the entire proximal tibia
section.

Following precontrast ?CT scans, 1.2 mL of Hexabrix (ioxaglate meglumine 39.3 % and
ioxaglate sodium 19.6 %; Mallinckrodt, St. Louis, MO, USA) was added to a 15 mL conical
tube and diluted with 1.8 mL of 0.1 M phosphate-buffered saline containing protease
inhibitors (1 % Protease Inhibitor Cocktail Set I, CalBiochem, San Diego, CA, USA),
yielding a 40 % solution of Hexabrix 39]. The tibia was then placed in this Hexabrix solution and was capped and incubated
in a covered, rocking water bath at 37 °C for 3 hours. After the incubation period,
the sample was removed, patted dry and placed in the plastic positioning device within
a ?CT holder containing a small amount of saline to help maintain sample hydration
during scanning. Postsoak scanning of the right tibia was performed in the same manner
as described above, except that the parameters were set differently to better visualize
the cartilage, with source energy of 55 kVp, 145 ?A at 8 W and an average scan time
of 42 minutes per sample 39], 40].

?CT evaluation of the cancellous bone at proximal tibial metaphysis

The cancellous bone compartment of the metaphysis was analyzed 1 mm below the growth
plate and extending 3 mm distally to include the metaphyseal secondary spongiosa.
Cancellous bone was evaluated in an ROI drawn on 100 consecutive slices, each 1.0 mm
in thickness, which best represented the central segment of the tibia 41], 42]. Cancellous bone parameters included bone mineral density (grams of calcium/bone
volume), tissue volume (bone and bone marrow), bone volume, bone volume/tissue volume
ratio, bone surface, bone surface/bone volume, trabecular number, trabecular thickness,
trabecular separation (distance between individual trabeculae), connectivity diameter
(connection between individual trabeculae), and structural model index for shape (Fig. 1a and b).

Fig. 1. a and b show two-dimensional micro-computed tomography (?CT) images of the proximal tibia
from sham control (Sham) (a) and medial meniscectomy vehicle-treated (MM?+?veh) rats (b). Red line indicated the area of cancellous bone evaluation at tibial epiphysis (e) and metaphysis
(m). Arrowhead points at the osteophyte. c and d show EPIC ?CT images of the same tibias depicted in (a) and (b). The red arrow indicates normal articular cartilage whereas the dotted arrow indicates osteoarthritic cartilage. e and f show larger EPIC ?CT images of the articular cartilage. The length of the medial
tibial plateau was measured for each sample and then divided into three zones ranging
from 0.8 to 1.0 mm in length. Zone 1 (Z1) was placed on the outside of the medial
edge of the joint and Zone 3 (Z3) on the inside of the tibial plateau adjacent to
the central collateral ligaments. Zones are delineated by dotted lines. The red arrow indicates normal articular cartilage whereas the dotted arrow indicates osteoarthritic cartilage

?CT evaluation of the cancellous bone at proximal tibial epiphysis

An ROI 2.0 mm?×?0.5 mm was drawn on the precontrast images to include only cancellous
subchondral bone underlying the articular cartilage. This region was drawn on the
100 consecutive slices (1.0 mm total thickness). Cancellous bone parameters included
bone mineral density (grams of calcium/bone volume), tissue volume (bone and bone
marrow), bone volume, bone volume/tissue volume ratio, bone surface, bone surface/bone
volume, trabecular number, trabecular thickness, trabecular separation (distance between
individual trabeculae), connectivity diameter (connection between individual trabeculae),
and structural model index for shape (Fig. 1a and b).

?CT evaluation of the epiphyseal cartilage

Using the postcontrast scans, contour lines were drawn around an ROI that included
the cartilage overlying the medial tibial plateau. The ROI was purposely drawn to
include a small amount of bone and soft tissue around the cartilage to ensure that
an appropriate threshold was selected to segment the cartilage from bone tissue and
soft tissue according to histographical analysis of the tissues. After the lower (70)
and the upper (440) threshold was determined, the contour lines were manually drawn
with semiautomatic contouring applied every 3–10 slices over a total of 300 slices
(3 mm) to capture most of the articular surface (Fig. 1c and d). The three-dimensional morphology of the entire articular cartilage layer drawn
was then visualized and quantified in terms of average cartilage thickness, volume
and surface area using direct distance transformation algorithms 43], 44].

Other ROIs were drawn and analyzed on this central midpoint of the articular surface,
corresponding to standard histological evaluation techniques for the articular cartilage
45]. The length of the medial articular cartilage was measured and divided into three
zones of equal length as already described for the subchondral bone. The parameters
of articular cartilage included cartilage volume and cartilage thickness (Fig. 1e and f).

Histology

After the completion of EPIC ?CT imaging of the articular cartilage, six tibias were
randomly chosen and placed in 10 % neutral buffered formalin for 72 hours prior to
demineralization for 8 days in Immunocal (Decal Chemical Corp., Tillman, NY, USA).
The tibias were then processed in paraffin and were serially sectioned at approximately
200 ?m intervals into 5 ?m-thick sections for staining. The slides were stained with
hematoxylin and eosin (HE) and toluidine blue for general structural evaluation and
with safranin O for the evaluation of cartilage damage. Thickness and degeneration
of the articular cartilage at the medial tibial plateau were determined on three longitudinal
sections of the proximal tibia using an ocular micrometer. Cartilage thickness was
measured separately on each of three zones, as suggested in the literature 45]–47]. The severity of OA lesions was graded on a scale adopted from Osteoarthritis Research
Society International (OARSI) histopathology instructions 45]. Five histological measures used in this study included damage score, osteophyte
size (?m), significant cartilage degeneration width (?m), cartilage thickness (?m)
and cartilage matrix loss width (?m).

Undecalcified bones were embedded in methyl methacrylate and were cut into 8 ?m-thick
longitudinal sections using a polycut sliding microtome (Leica Biosystems, Nussloch,
Germany) or into 20 ?m-thick sections using a bone cutting and grinding system (Exakt
Norderstedt, Germany). The unstained sections were used first to assess new bone formation
at the medial tibial epiphysis, and these sections were subsequently stained according
to Von Kossa’s method to quantify mineralization.

Statistical analysis

Data are given as means?±?standard deviations (SDs). Differences were tested for significance
using three-factor repeated-measures analysis of variance (ANOVA) with interactions
(Sigma Plot, version 12.2, Systat Software, Chicago, IL, USA). Post hoc comparisons
of means with a Bonferroni correction for multiple comparisons were performed only
when interaction effects were significant. p values less than 0.05 were considered statistically significant.