Ankylosing spondylitis (AS) is a common rheumatic disease with a prevalence of 0.9% worldwide.1 The mechanism of pathogenesis is poorly understood despite the recognition of the role of the human leukocyte antigen B27 (HLA-B27) in 90-95% of AS patients, and the interleukin 23 receptor (IL-23R) polymorphism.2,3 AS involves both inflammatory erosive osteopenia and unusual bony overgrowth gradually replacing joints tissue.4 When lesions occur in the spine, the ultimate bone remodeling occurs with complete vertebrae fusion. The delay between onset of symptoms and diagnosis can reach up to 10 years,5 incurring significant disability and economic cost.6
Currently, the most effective medication for AS encompasses biologic agents blocking TNF-α.7,8 The ability of these agents to prevent ankyloses remains controversial, despite the fact that recent long-term studies provided encouraging results.8,9 The development of different animal models have been useful for the study of AS pathogenesis and for testing of new therapeutic concepts. Progressive ankyloses in ank/ank mice, HLA-B27 transgenic mice and rats, as well as challenges with complete or incomplete Freund’s adjuvant associated with proteoglycans have been explored.10-14
Imaging techniques aim at bridging preclinical and clinical studies. It has been shown that computerized tomography (CT) and magnetic resonance imaging (MRI) are better techniques for the detection of earlier AS features in humans than radiography.15-17 We applied MRI to follow longitudinally in vivo, at the level of the rat spine, changes induced by complete Freund’s adjuvant (CFA) inoculation into the tail.18,19 Post-mortem analyses included micro-CT imaging and histology of the isolated spine. This study shows interesting findings in imaging readouts and reiterate the fundamental role played by histology in order to validate non-invasive techniques.
Materials and Methods
Experiments were performed according to animal license number BS-1505, granted by the Veterinary Authority of the City of Basel, Switzerland.
Animals and disease induction
Han Wistar rats (n=2, Charles River Laboratories, Wilmington, MA, USA) were acclimatized in a controlled environment for one week. Polyarthritis was induced in one rat by an intracutaneous injection at the tail base, of 0.1 mL of CFA [6 mg Mycobacterium tuberculosis (H37RA, DIFCO, Becton Dickinson) suspended in paraffin oil (1 mL)].20 The other animal served as control. Animals were imaged from the second week following CFA injection, when the paws were already inflamed. The study was finished on day 60 after CFA.
Magnetic resonance imaging
For MRI, rats were anesthetized with 2.0% isoflurane (Abbott, Cham, Switzerland) in a mixture of O2/N2O (2:1), administered via a nose cone. Body temperature was kept at 37±1°C using a heating pad. Measurements were carried out with a Biospec 70/30 spectrometer (Bruker Medical Systems, Ettlingen, Germany) operating at 7.0 T. A quadrature resonator with 70 mm inner diameter (Bruker) was used for signal excitation and detection. Three-dimensional (3D) gradient-echo imaging (see Figure 2 legend for parameters) was started 6 min after i.v. injection of Gd-DOTA (Guerbet, Aulnay-sous-Bois, France; 0.15 mL) as a bolus into the tail vein.21
At the end of the study animals were euthanized by an overdose of thiopental (Pentothal®, Abbott; 250 mg/kg i.p., 0.2 mL) and the spines were excised for micro-CT and histological examination. Spine samples were first fixed in 4% neutral buffered paraformaldehyde for two days and then kept in 70% ethanol at 4°C until completion of micro-CT.
Micro computerized tomography
Scans of fixed spine were performed on a μCT40 instrument (Scanco Medical, Brüttisellen, Switzerland) using the following parameters: energy/intensity: 70 kVp, 57 µA, 4 W; image matrix: 2048x2048 pixels, voxel size: 20 µm. Segmentation thresholds of 190 and 280 and a Gaussian filter (σ = 0.8, support of one voxel) were used for image analysis.
Following micro-CT imaging, samples were kept in a decalcification solution (ImmunoCal™, Decal Chemical Corporation, Suffern, NY) for 14 days until decalcification was complete as verified chemically using a procedure previously described.22 After sample dehydration and paraffin embedding, 5-µm thick sections were cut and stained with Hematoxylin and Eosin (HE). Subsequently, additional stainings were performed to distinguish cartilage, bone, muscle, fibrotic tissue and monocytes/osteoclasts. Cartilage was stained by the Safranin O/Fast green procedure adapted from Kalscheur.23 Safranin-O stained the cartilaginous areas until the tide-mark, while Fast green stained the subchondral mineralized bone. The Aldehyde-Fuchsin-Picrosirius red (AFPS) staining was used to differentiate cartilage, collagens or bone. Under this specific staining, newly formed bone has the aspect of woven bone and polarizes in green rather than orange/red.24 Osteoclasts were detected with an anti-CD68 antibody applied on paraffin sections.25,26 After an antigen retrieval step in a 10 mM citrate buffer (pH 6) at 98°C, the primary monoclonal antibody anti-rat ED1 (MCA341R, Bio-Rad, Puchheim, Germany) was incubated at 0.05 μg/mL followed by a biotinylated secondary antibody (BA-2001, Vector Lab., Burlingame, CA) and amplified by the Avidin Biotin - HRP complex (Elite PK-6100, Vector Lab.) according to manufacturer’s recommendations. After the reaction was revealed with DAB, sections were counterstained with Mayer’s Hematoxylin, washed, dehydrated and mounted with Pertex.
The main features of AS were observed by histology. Globally, the L5-L6 lumbar spine of the CFA rat (Figure 1 A-b) was severely affected as demonstrated by the damage of the vertebral bone. Subchondral tissues were infiltrated by immune cells like osteoclasts responsible for bone resorption (Figure 1 A-c), polymorphonuclear cells (Figure 1 A-d), and lymphocytes (Figure 1 A-e). Figure 1 B-a shows the invasion of joint space by a synovial pannus (sp) resulting from fibroblast-like synoviocyte proliferation, leading to cartilage destruction (ca) and bone invasion (bo). The remaining nucleus pulposus (np) of intervertebral disc surrounded newly-formed bone (bo) replacing the annulus fibrosus (Figure 1 B-b). No osteoclasts were observed in this newly-formed bone (data not shown). Figure 1 B-c displays the bony ankylosis process resulting in a newly formed bone, so-called woven bone (bo), well-observed under polarized light. This confirmed the presence of inflammation and erosion, as well as new bone formation at the levels L5-L6 of the spine of the CFA-rat.
MRI images of the longitudinal development of injury induced by CFA inoculation at the lumbar levels L5-L6 of the spine are summarized in Figure 2. High intensity signals (red arrows) reflecting inflammation have been detected throughout the study. At day 14, the inflammatory response occurred along ligaments. It expanded to nearby soft tissues and also to regions within the eroded bone (magenta arrows) at later time points. Moreover, images indicate the formation of bone-like structures (green arrows) from day 27 onwards.
Figure 3A presents MRI images of the lumbar spine at day 53 post-CFA, showing inflamed areas (characterized by high intensity signals) around the levels L5-L6 of the spine (red arrows) and within the bone (magenta arrows), as well as the formation of new bony structures (green arrows). Images from the control rat displayed none of these features. Post-mortem micro-CT showed severe axial bone remodeling (erosion and new bone formation) and beginning dorsal spinal fusion in the lower lumbar spine region of the CFA rat (Figure 3B). An excellent agreement was found between the new bone formation as revealed histologically and the formation of bone-like structures as indicated by in vivo MRI (Figure 3C).
The histological data presented here indicate the potential of the rat adjuvant model to mimic some of the important human pathological features of AS, such as inflammatory response showing polymorphonuclear cells and lymphocytes as well as subchondral tissue infiltration by osteoclasts leading to bone resorption, and the disappearance of intervertebral discs and their replacement by a well-vascularized fibrotic tissue followed by ossification of fibrocartilage in the spine. This bone formation is one of the early signs of the bony fusion process of vertebrae. Using MRI we could visualize progressive structural damages, such as edema and inflammatory lesions as well as bone erosion and formation in the lumbar spine over time. Post-mortem micro-CT confirmed this bone remodeling.
In this pilot study, only one sampling time was used to compare MRI, micro-CT and histology. MRI showed the inflammatory response occurring along ligaments, then expanding to soft tissues and bone, and the formation of bone-like structures over time, demonstrating the progressive nature of this pathology. Even if more extensive studies are needed to investigate the relationship between inflammation in the spine and bone remodeling, the development of progressive damage observed in vivo indicated a correlation with histopathological features of AS.
The present study including three complementary approaches clearly demonstrates the potential of histology in the demonstration of pathologic features provided by non-invasive imaging. Moreover, our data suggest that in vivo imaging in this AS model may be used to validate the use of imaging in translational activities as therapies are moved into patient studies. Since AS has a very long disease development in humans,27 having access to well-characterized imaging readouts for early diagnosis becomes increasingly essential for therapy intervention studies.28,29