The goal of this protocol is to generate a nude rat osteoporosis-related vertebral compression fracture model that can be longitudinally evaluated in vivo using a semiautomated microcomputed tomography-based quantitative structural analysis.
Osteoporosis-related vertebral compression fractures (OVCFs) are a common and clinically unmet need with increasing prevalence as the world population ages. Animal OVCF models are essential to the preclinical development of translational tissue engineering strategies. While a number of models currently exist, this protocol describes an optimized method for inducing multiple highly reproducible vertebral defects in a single nude rat. A novel longitudinal semiautomated microcomputed tomography (µCT)-based quantitative structural analysis of the vertebral defects is also detailed. Briefly, rats were imaged at multiple time points post-op. The day 1 scan was reoriented to a standard position, and a standard volume of interest was defined. Subsequent µCT scans of each rat were automatically registered to the day 1 scan so the same volume of interest was then analyzed to assess for new bone formation. This versatile approach can be adapted to a variety of other models where longitudinal imaging-based analysis could benefit from precise 3D semiautomated alignment. Taken together, this protocol describes a readily quantifiable and easily reproducible system for osteoporosis and bone research. The suggested protocol takes 4 months to induce osteoporosis in nude ovariectomized rats and between 2.7 and 4 h to generate, image, and analyze two vertebral defects, depending on tissue size and equipment.
More than 200 million people worldwide suffer from osteoporosis1. The underlying pathological decrease in bone mineral density (BMD) and altered bone microarchitecture increase bone fragility and, consequently, the relative risk of fracture2. Osteoporosis is so prevalent and detrimental to health that the WHO has defined it a major public health concern. Furthermore, as the world's population is expected to age, osteoporosis is expected to become even more common.
Osteoporotic vertebral compression fractures are the most common fragility fractures, estimated at more than 750,000 a year in the US. They are associated with significant morbidity and as much as a nine-times higher mortality rate3. In clinical trials, currently available surgical interventions, such as vertebroplasty and kyphoplasty, were found to be no more effective than a sham treatment4,5, leaving only pain management available to these patients. Since current OVCF treatments are limited, it is imperative to develop an animal model that can replicate the disorder6,7,8. Such animal models could facilitate both the investigation of current treatment methods and the development of novel therapies that will translate into clinical practice. Osteoporosis has been induced and sustained in model animals through the administration of a low-calcium diet (LCD) in conjunction with ovariectomy1,9,10,11,12,13,14,15. To further model the bone loss associated with OVCFs, vertebral bone defects were established in osteoporotic immunocompetent rats 16,17,18,19,20,21,22,23,24. In this work, a vertebral defect model of immunocompromised rats with modeled osteoporosis is presented. This novel model can be used to assess cell-based therapies involving stem cells derived from various sources and species for the repair of challenging fractures, such as OVCFs.
Bone imaging is a crucial part of the evaluation of fractures and bone diseases. Advanced imaging methods were developed for the accurate assessment of structural bone changes and regeneration strategies25. Among them, µCT imaging has emerged as a non-invasive, easy-to-use, and inexpensive method that provides high-resolution 3D images. µCT imaging has several advantages over other modalities in evaluating osteoporosis patients, as it offers high-resolution 3D bone microarchitecture26 that can then be quantitatively analyzed. The latter can then be used to compare the therapeutic effects of proposed treatments. Indeed, in vivo µCT imaging is a gold standard for vertebral defect regeneration monitoring1,16,27. However, few publications28,29,30,31 have employed automated registration tools to minimize the user-dependency, interpolation bias, and precision error of µCT imaging-based analysis. Recently, we were the first to use a registration procedure to improve the analysis of bone regeneration in a standardized bone void, as explained in this protocol32 .
The method described here can be used to study the effect of novel cell therapies for OVCFs, unhindered by host T-cell responses that might reject xenogeneic or allogeneic cells. Osteoporosis is induced in young rats through ovariectomy (OVX) and 4 months of an LCD. The young age of the OVX rats, combined with the LCD allowed, us to reach a low peak bone mass, mimicking postmenopausal osteoporosis by leading to irreversible bone loss. This can be explained partly by the fact that, during the LCD and at around 3 months of age, the rats transition from the bone modeling to remodeling phase at the lumbar vertebrae33, thereby increasing the likelihood of maintaining the osteoporosis over time. Using young animals makes this model more cost effective, as they cost less. Nonetheless, it is limited by inherently not accounting for the biological changes in the aging animal.
All animal experiments were performed under a protocol approved by the Institutional Animal Care and Use Committee (IACUC) of Cedars-Sinai Medical Center (Protocol # 3609). Anesthesia was administered for all imaging and surgical procedures. All animals were housed in accordance with approved IACUC protocols.
NOTE: The experimental design of this protocol is shown in Figure 1. Purchase six-week-old rats with their ovaries surgically removed and feed them an LCD consisting of 0.01% calcium and 0.77% phosphate. After a period of 4 months of an LCD, drill a critical-size vertebral defect in the fourth and fifth lumbar vertebral bodies (L4-L5). Following surgery, image the rats on day 1 and weeks 2, 4, 8, and 12 after defect establishment. Locate defect margins on the day 1 scan, reorient to a standard position, and define a cylindrical volume of interest (VOI). Automatically register the subsequent µCT scans (i.e., for weeks 2, 4, 8, and 12) of each rat to the standard position defined for the corresponding day 1 scan. Apply the day 1 predefined VOI to the registered scans. Assess the bone volume density and apparent density of the VOIs.
1. Induction of Osteoporosis
2. Vertebral Defect Model
NOTE: The timing is 40 – 50 min per animal.
3. MicroCT Scanning
NOTE: The timing is 30 – 40 min per animal.
4. Vertebral Separation
NOTE: The timing is 20 – 30 min per sample.
5. Definition of the VOI for Longitudinal Quantitative Evaluation
NOTE: The following steps depend on whether the scan is from day 1 after surgery (reference vertebra) or from the subsequent time points (target vertebrae).
6. MicroCT Analysis
NOTE: The timing is 10 – 20 min per sample.
7. Euthanasia
Using this protocol, one can image and quantify the regeneration of n = 8 modeled osteoporotic vertebral defects across different time points. The anatomic match obtained by the registration procedure allows for the analysis of the same VOI at all time points. This results in a highly accurate longitudinal 3D histomorphometric analysis, even when the margins of the original defect are no longer recognizable. We used five time points (day 1, week 2, week 4, week 8, and week 12) as an example for the longitudinal evaluation of bone regeneration (Figure 7). Regeneration can be evaluated both by the qualitative assessment of 2D cross-sections and 3D images (as illustrated in Figure 7A) and by the quantitative comparison of the bone quantity (BVD) and quality (AD) (Figure 7B). The following morphometric indices can be determined for newly formed bone: (i) TV, including bone and soft-tissue volumes (TV, mm3); (ii) volume of mineralized tissue (BV, mm3); (iii) bone volume density (BV/TV); and (iv) bone mineral density (BMD, mg hydroxyapatite per cm3). Specifically, minimal bone formation (5% increase in bone volume density) was observed 2 weeks after defect establishment. After two weeks, no significant differences in bone formation were observed when compared to later time points. Overall, although there was some degree of bone formation, which peaked at approximately 10% by Week 8, it was minimal enough to maintain the bone void over time.
Figure 1: Protocol Design. The key steps in the protocol are outlined. First, ovariectomized nude rats subjected to four months of a low calcium diet (LCD) were operated upon to create standard critical-sized defects in two lumbar vertebral bodies. The rats were imaged on day 1 and weeks 2, 4, 8, and 12 post-op. The day 1 scan was reoriented to a standard position, and a cylindrical VOI was defined using the defect margins. Subsequent µCT scans of each rat were automatically registered to the standard position defined for the corresponding day 1 scan. The day 1 predefined VOI was then applied to the registered scans. The bone volume density and apparent density of the VOI were used to assess new bone formation. Please click here to view a larger version of this figure.
Figure 2: Vertebral Defect Surgery. The key steps in the surgical generation of vertebral defects are illustrated. First, rats were placed on a heating pad (A). A midline incision was made through the skin (B) and then the linea alba (C) to expose the abdominal cavity (D). The intestines were reflected to expose the posterior abdominal wall (E), and the lumbar spine was exposed using thermocautery (arrow, F-G). Defects were drilled in the fourth (H, arrow pointing to the drill; I, arrow pointing to the defect) and fifth (J, arrows pointing to defects) lumbar vertebral bodies. Finally, the aponeurosis (K) and skin (L) were sutured. Please click here to view a larger version of this figure.
Figure 3: Vertebra Separation. The key steps in the contouring of a vertebra of interest are shown. (A-I) Contoured (green line) representative 2D slices along the length axis of a vertebra are shown. A 3D representation of the full spine (J) can be compared to the separated vertebra (K). Please click here to view a larger version of this figure.
Figure 4: Reference Vertebra Positioning. Representative slices in two planes are shown of a vertebra before and after rotation to a standard position. First, using a representative XY-slice (A), the angle (B, green) needed to rotate the defect (B, red square) to become parallel to the Y-axis (B, yellow) is determined and then used to create the rotated image (C). Then, using a representative YZ-slice (D), the angle (E, green) needed to rotate the defect (E, red square) to become parallel to the Z-axis (E, yellow) is determined and then used to create the rotated image (F). Please click here to view a larger version of this figure.
Figure 5: Target Vertebra Registration. Representative slices at three planes of the target vertebra (marked in green) and reference vertebra (marked in red) before (A-C) and after (D-E) registration are shown. Note the yellow color, indicating overlap between target and reference vertebrae, and the white arrows that point to green bone after regeneration, indicating bone formation. Please click here to view a larger version of this figure.
Figure 6: VOI Analysis. Representative slices in two planes with the contoured volume of interest are shown. A circular contour is positioned at the center of the defect in a representative ZX-slice (A). After contouring all ZX-slices, the complete defect volume can be seen in the XY-plane (B). Please click here to view a larger version of this figure.
Figure 7: Longitudinal Analysis of Vertebral Defect Regeneration. Qualitative and quantitative representative bone regeneration analysis results are shown. (A) A representative vertebral defect at various time points is depicted in each panel as a frontal 3D image (top panel) with bone formation in the void indicated in red, a sagittal 2D image (middle panel), and an axial 2D image (bottom panel). The quantitative analysis of bone formation in the voids was performed. Bone volume density (B) and apparent density (C) were calculated and compared using a repeated measures two-way ANOVA with Bonferroni correction for multiple comparisons. The error bars represent SEM. **** – p<0.0001. Please click here to view a larger version of this figure.
Steps | Problem | Possible reason | Solution |
2.3 | Animal gasping under anesthesia | Excess isoflurane delivery | Reduce the concentration of isoflurane delivered to the animal. |
Animal responds to toe pinch | Insufficient isoflurane delivery | Increase the concentration of isoflurane. | |
2.7-2.12 | Heavy bleeding | Vascular damage | Use a sterile cotton swab to apply pressure or cautery to stop bleeding. |
The animal has difficulty breathing | The diaphragm was punctured | Euthanize the animal to prevent suffocation. | |
Leakage of intestinal contents | The gastrointestinal tract was punctured | Euthanize the animal to prevent further complications. Prevent it by lifting the aponeurosis away from underlying intestines before cutting. | |
Blood emerges from the drilling site | A blood vessel was punctured | Apply a sterile cotton swab until bleeding stops. | |
Animal suddenly shakes during drilling | The drill went too deep and damaged the spinal cord | Euthanize the animal to prevent further complications. | |
The bone defect looks incomplete | The drill didn’t go deep enough | Reposition the drill head inside the defect and drill deeper | |
2.15-2.24 | Suture breaks | The suture was pulled too tightly | Replace the entire suture. If breaking occurs often, use a size thicker suture. |
Animal is slow to recover from anesthesia | The animal is hypothermic | Increase the temperature of the heating pad or apply an additional source of heating (e.g. heating lamp). | |
Sutures are open | The sutures were placed loosely, or the animal did strenuous activity | Reapply the sutures and apply Dermabond directly to the sutures and between them. | |
3 | Scanned image appears with low resolution, noisy or scattered | Scanning parameters need to be adjusted | Adjust the parameters of the scanning protocol. Refer to Bouxsein et al. for more guidelines for scanning. |
Scanned image appears blurry | The animal moved during the scanning process | Rescan the animal. If movement continues, increase isoflurane concentration. | |
5 | The registration of target vertebra wasn’t successful | Vertebral separation wasn’t done properly | Recontour the vertebra: Make sure all parts of vertebra are included, and exclude any adjacent structures. |
Large difference in positioning of vertebrae | Reposition the target vertebra into the same orientation as the reference vertebra using rotations and flip (Step 29A). | ||
Analyze can’t properly recognize the bone structures | Apply a threshold in the registration module to remove the background noise from bone samples. | ||
The registered vertebrae are different | Create 3d images of your samples and match the correct vertebrae across the different time points. | ||
6 | The total volume (TV) is different between samples | Either a different numbers of slices or a different contour was used | Make sure to always use the same contour size and the same number of slices. |
Bone mineral density (BMD) value is abnormal | Inadequate calibration of microCT | Calibrate the microCT for correct hydroxyapatite standards |
Table 1: Troubleshooting. Potential problems and solutions are presented for different steps in the protocol.
Osteoporosis is the most prevalent cause of vertebral compression fractures caused by an increased load on the spine and that result in the collapse of the vertebral body. However, it is practically impossible to generate an injury in a rodent that authentically replicates a similar vertebral collapse. Instead, researchers create a cylindrical void in the center of the vertebral body to mimic OVCFs16,17,18,19,20,21,24,38,39. Since there is no consistency in the literature in terms of defect size, a critical-sized defect was defined as one that does not spontaneously heal fully without an intervention within 3 months post-op16,17.
Although the method of combining ovariectomy with an LCD to rapidly induce osteoporosis was previously published1,13, we were the first to show that applying this approach to athymic rats results in an efficient, rapid, and irreversible decrease in vertebral trabecular bone volume and mineral density40. This is a reproducible small-animal model that is unhindered by the rodent immune system and that does not have a need for added immunosuppression, as used by others24.
Our surgical protocol generated multiple identical critical lumbar vertebral defects40. This results in highly consistent and easily comparable and quantifiable defects across animals. We believe that defects produced using this approach are superior to vertebral defect models generated in caudal vertebrae1,19,41 because the rat tail is subjected to biomechanical forces that are significantly different from those involving the rat lumbar spine.
Critical steps within this protocol include avoiding intra-operative hypothermia and taking caution when drilling the fragile vertebrae of ovariectomized nude rats after an LCD. After generating the vertebral defect, it is monitored via a temporal sequence of in vivo µCT scans at set time points for the longitudinal assessment of bone repair. Maintaining the same scan settings is critical. The vertebrae are then contoured and separated from the rest of the scan. Contouring an identical total volume for all scans of a vertebra and avoiding grayscale value changes are critical. A commercially available multiple image registration algorithm facilitates the extraction of anatomically corresponding baseline VOIs to all subsequent time points. Finally, these VOIs are analyzed for bone volume, apparent density, etc. It is critical to analyze all VOIs using the same parameters. This technique provides a highly accurate and straightforward longitudinal 3D µCT analysis that is not user-dependent.
This method could be applied to any longitudinal bone defect regeneration analysis. The vertebral defect model used here is a convenient model for this application, as its bone structure is unique and can be easily registered to the same anatomical position. However, any bone regeneration could be analyzed under the same conditions by properly separating the same bone of interest throughout the longitudinal scans. It is imperative to include separated bone samples with the same anatomical features. This potential problem and others are described in Table 1, along with possible reasons and suggested solutions. The anatomic match obtained by the registration procedure can only occur if the samples include the same anatomical features. The registration will allow the user to apply the exact predefined VOI of the first scan to all remaining time points, resulting in a highly accurate 3D histomorphometric analysis over time. Bone volume density and apparent density of the VOI can be used to assess new bone formation.
While potentially widely applicable, the model presented here is not without limitations. The use of athymic nude rats could be considered a limitation, as it could potentially mask some immune-mediated processes that may be of importance to regeneration. Second, modeling osteoporosis through a combination of ovariectomy and an LCD in young rats, as previously published1,13, is limited in its ability to mimic the biology of the elderly patient population. Third, OVCFs were modeled by a surgical procedure, as the only other animals to have osteoporosis-related fractures are primates42. Finally, while the rat lumbar spine is the best available model for the human lumbar spine—where most vertebral fractures develop—the lack of axial weight bearing in the rodent spine is also a limitation.
This protocol is modular and therefore could be easily modified to the researcher's needs. For example, the athymic ovariectomized rats could be used to study other osteoporosis-related fractures. Should a researcher choose to use our approach to semiautomated bone regeneration analysis, it could be applied to any fracture model using longitudinal structural imaging, not necessarily micro-computed tomography. Furthermore, additional information could be gathered by simultaneously using additional imaging modalities such as magnetic resonance imaging.
The OVCF model presented in this protocol could be used to study novel therapeutic approaches to this clinically unmet need. Furthermore, our semiautomated analysis approach can be successfully used to perform a similar analysis that is less user-dependent and provides better accuracy than other methods16. Particularly noteworthy is the fact that we used commercially available visualization and analysis software that can be used by any researcher—software that supports additional imaging modalities, such as magnetic resonance imaging and nuclear imaging. Therefore, we believe that this method is highly generalizable and is only limited by the availability of in vivo imaging capabilities and registration software.
The authors have nothing to disclose.
The research was supported by a grant from the California Institute for Regenerative Medicine (CIRM) (TR2-01780).
Isoflurane | MWI Animal Health, Pasadena, CA | 501017 | |
BetadineSolution | MWI Animal Health, Pasadena, CA | 4677 | |
Chlorhexidine Gluconate2%scrub | MWI Animal Health, Pasadena, CA | 510083 | |
Isopropyl Alcohol 70%-quart | MWI Animal Health, Pasadena, CA | 501044 | |
Carprofen | MWI Animal Health, Pasadena, CA | 26357 | |
Buprenorphine0.3mg/mL | MWI Animal Health, Pasadena, CA | 56163 | |
Ovariectomized Athymic nude rats | Harlan Laboratories, Indianapolis, IN | Hsd:RH-Foxn1 rnu | |
Low calcium food | Newco Distributors, Inc., CA | 1814948 (5AV8 AIN-93M w/low calcium) | |
Phosphate Buffered Saline | Life Technologies Corporation | 14190250 | |
Dermabond | J AND J ETHICON | DHVM12 | |
Anesthesia machine | Patterson Scientific | TEC 3EX | |
Slide Top Induction Chambers | Patterson Scientific | 78917833 | |
ProStation Heated Workstation | Patterson Scientific | 78914731 | |
Surgical drape | HALYARD HEALTH INC | 89101 | |
Magnetic fixator retraction system | Fine Science Tools, Inc., CA | 18200-50 | |
Dissecting Scissors, 10cm, Curved, SS | World Precision Instruments, FL | 14394 | |
Iris Scissors, 11.5cm, 45°Angle, Serrated, Sharp/Sharp | World Precision Instruments, FL | 503225 | |
Forceps, no. 5 | World Precision Instruments, FL | 555048FT | |
Micro Mosquito Hemostatic Forceps | World Precision Instruments, FL | 503360 | |
Sterile cotton gauze | Medtronic, MINNEAPOLIS, MN | 9024 | |
Absorption Spears – Mounted/Sterile | Fine Science Tools, CA | 18105-01 | |
Syringe, 1 ml | TERUMO TERUMO MED | SS-01T | |
Needle, 25gauge | BD MED SYS INJECTION SYS | 305127 | |
Laminar flow hood | Baker | SterilGARD e3-Class II Type A2 Biosafety Cabinet | |
Thermal Cautery Unit | World Precision Instruments, FL | 501292 | |
Micro-Drill OmniDrill115/230V | World Precision Instruments, FL | 503598 | |
Trephines for Micro Drill, 2mm diameter | Fine Science Tools, CA | 18004-20 | |
3-0 Vicryl undyed 27” SH taper | J AND J ETHICON | 1663G | |
4-0 Ethilon black 18” PC3 conventional cutting | J AND J ETHICON | 1954G | |
Conebeam in vivo microCT (vivaCT 40) | Scanco Medical | vivaCT 40 | |
SCANCO Medical microCT systems software suite | Scanco Medical | vivaCT 40 | |
Analyze software | Biomedical Imaging, Mayo Clinic, Rochester, MN | Analyze 12 | Image analysis software |
Veterenery eye ointment |