Mouse Model of Fracture Pain

This study, including care of the animals involved, was conducted according to the official edict presented by the French Ministry of Agriculture (Paris, France) and the recommendations of the Declaration of Helsinki. Therefore, these experiments were conducted in an authorized laboratory and under the supervision of an authorized researcher (I.T.). These experiments were approved by our institutional animal care and use committee, and this study was conducted in accordance with the International Association for the Study of Pain Guidelines on the Use of Animals in Experimental Research.15Adult C57 BL/6 male mice (The Jackson Laboratories, Bar Harbor, ME) were used in all experiments. The animals were housed individually in isolator cages with solid floor covered with 3 cm of soft bedding and were fed and watered ad libitum . Animals were on a 12-h light–dark cycle.

All mice were anesthetized with 1.5–2% sevoflurane delivered via  cone nose. After antiseptic preparation of the right paw with povidone iodine, a unilateral, closed fracture was produced in the right tibia using a specially designed fracture apparatus (blunt guillotine). The fracture apparatus consists of four parts: a frame, an animal support system, a guillotine ramming system, and a 300-g weight. The support anvil was made with an adjustable foot rest that ensures that all of the fractures are at the same level by positioning the mouse leg on the anvil with the foot against the foot rest. For the intramedullary pinning using a sterile technique, a hole was made above the tibial tuberosity percutaneously using a 27-gauge needle (Becton Dickinson and Company, Drogheda, Ireland). Then the needle was directed directly into the medullary canal. By rotating the needle, the canal was reamed to 5 mm up to the ankle joint. The end of the needle was cut as short as possible so that the skin could roll over and cover it. No suture was used.

Then, the mouse was placed with the leg on the anvil so that the blunt guillotine lined up with the proximal third of the tibia. The 300-g weight was dropped from a height of 9–10 cm, fracturing the tibia shaft. Other heights were tested, but this one produced the best results. Radiography confirmed the fracture.

In group 1 (sham incision group, n = 10) mice were anesthetized, and a skin puncture of the knee was performed using the same 27-gauge needle.

In group 2 (sham pinning group, n = 10) mice were anesthetized, and an intramedullary pinning was placed as described above, but no fracture was performed.

In group 3 (fracture group, n = 10) mice were anesthetized, an intramedullary pinning was placed, and a closed fracture was performed as described above.

All mice of each group underwent postoperative pain testing. The animals were tested before surgery (baseline); 2, 4, and 6 h after surgery; once daily during the first 7 postoperative days; and the week after surgery.

The aim of this second part was to assess the effect of systemic morphine administration on enhanced responses to postoperative testing.

Forty other mice were randomly assigned to receive morphine (3 or 10 mg/kg body weight), ketoprofen (50 mg/kg body weight), or placebo (vehicle). Surgery was performed as described above. Testing (mechanical stimulation, hot plate test, and pain rating scale) were performed before the surgery and 2 h after the surgery (T0 = baseline). Then mice were randomly assigned to receive 3 mg/kg morphine (n = 10), 10 mg/kg morphine (n = 10), ketoprofen (n = 10), or a saline vehicle subcutaneously (n = 10). Responses to mechanical and heat stimuli and pain rating scale score were determined each 30 min after T0 until 240 min after the administration. The effects of morphine administration and of ketoprofen administration were also assessed on behavioral testing on postoperative days 1 and 2 using the same procedure. Experiments were conducted following a double-blind protocol.

Three tests were used to assess pain behavior: (1) mechanical nociception assessed by the withdrawal response to von Frey filament application, (2) thermal nociception assessed by the withdrawal response to thermal stimulus (hot plate test), and (3) subjective pain determined using a pain rating scale as described by Attal et al.  16 

Unrestrained mice were placed beneath a clear plastic chamber on an elevated mesh floor and allowed to acclimate. Withdrawal responses to mechanical stimulation were determined using calibrated von Frey filaments applied from underneath the cage through openings in the plastic mesh floor against the hind paw plantar skin at approximately the middle of the paw at the fractured side. The filament was pushed until it slightly bowed and then it was jiggled in that position for 6 s. Each von Frey filament was applied once, starting with 0.008 g and continuing until a withdrawal response was reached that was considered a positive response. After a 5- to 10-min rest period, each filament was again applied once, beginning with 0.008 g until a withdrawal response was elicited. This was repeated a third time 5–10 min later. The lowest force from the three tests producing a response was considered the withdrawal threshold.

Thermal nociception was measured by a modified hot plate test.17The time that a mouse would leave its hind paw on a hot plate at 52°C reflects thermal nociception (thermal latency). The paw was removed from the plate after a maximal time of 12 s by the investigator to avoid thermal injury and thermal hyperalgesia.17This test was repeated three times on each hind paw for each mouse.

A subjective pain rating scale (0–5) modified from that described by Attal et al.  16was used to quantify the pain, where 0 is normal, 1 is curling of the toes, 2 is eversion of the paw, 3 is partial weight bearing, 4 is non–weight bearing and guarding, and 5 is avoidance of any contact with the hind limb.

The results of behavioral testing were not normally distributed and thus were analyzed nonparametrically. To assess whether the withdrawal responses changed over time, the Friedman test was used. When the Friedman test was significant (P < 0.05), pairwise comparisons were made using the Wilcoxon signed rank test. Time point comparisons between groups were made using first a nonparametric Kruskal-Wallis test. When the Kruskal-Wallis test was significant (P < 0.05), pairwise comparisons were made using the Mann–Whitney U test. The effect sizes were also estimated using Cohen d values. Cohen d values are considered to be a small effect size at 0.2, a moderate effect size at 0.5, and a large effect size at greater than 0.8.

Throughout the experimental period, all mice remained well groomed and maintained normal food and water intake. No signs of spontaneous pain behavior, such as licking, biting, or flinching, were noticed after the surgery. The experiments were conducted following a double-blind procedure because neither swelling nor hematoma occurred after fracture.

No difference in the measured parameter was observed in any of the groups at baseline.

Thermal nociception, mechanical nociception, and subjective pain were significantly modified in group 3 compared with control groups 1 and 2.

As shown in figure 1, withdrawal response to mechanical stimulus decreased in group 3, indicating an increased nociception, whereas no change was observed in groups 1 and 2.

The same pattern was observed in figure 2; the response latency reflecting an increased nociception was significantly decreased in group 3, whereas no change was observed in groups 1 and 2. Effect sizes (Cohen d) for mechanical and thermal stimulation are shown in tables 1 and 2, respectively.

The subjective pain scale shown in figure 3was significantly increased in group 3 compared with groups 1 and 2, which remained steady.

When tests were repeated for 240 min in morphine- or ketoprofen-treated animals, reduction of mechanical nociception, thermal nociception, and subjective pain scale score were observed.

As shown in figures 4 and 5, withdrawal response and thermal response increased transiently after a single morphine injection. Moreover, the increase was dose dependent. As shown in figures 4 and 5, withdrawal response and thermal response increased transiently after a single ketoprofen injection. Effect sizes (Cohen d) for mechanical and thermal stimulation are shown in tables 3 and 4, respectively. The subjective pain scale (fig. 6) was transiently reduced in a dose-dependent manner after morphine injection and returned to maximal value after 240 min. It was also transiently reduced after ketoprofen injection (fig. 6). The effects of morphine administration and of ketoprofen administration were identical on behavioral testing on postoperative days 1 and 2, as shown in insets of figures 4–6.

The current study developed a method of closed fracture of the tibia shaft to validate a model of postfracture bone pain in mice. Because no change was observed in the control groups (groups 1 and 2), it can be reasonably assumed that the effects observed in the fractured group (group 3) result from events occurring at the site of the bone lesion. Moreover, this model is morphine– and nonsteroidal antiinflammatory drug (NSAID)–sensitive, a situation closely related to clinical practice.

To our knowledge, this is the first report of posttraumatic model in mice. However, one model after bone injury has been recently reported in rats.18Besides the considerable advantage of using a mouse model versus  a rat model (genetically modified mice), the described rat model was different. In this model, bone lesions were induced by bone holing after skin incision, thereby resulting in several origins of pain mechanism (bone, skin, muscles, and so on).

The mouse model of standard closed tibia fracture described here has been adapted from a rat model of closed tibia fracture.10,12However, this model was only used by orthopedists to assess bone reconstruction.13,14In a recent model described by Bonnarens et al. , closed femur fracture was used in mice,9,11mainly for evaluation of bone reconstruction. They used only a simple subjective pain evaluation.19,20However, in our model, the tibia (and not the femur) was fractured, producing a pain mainly due to bone fracture (and probably less to hematoma and inflammation) and allowing us more easily, because of its peripheral localization in the limb, to accurately evaluate pain, i.e. , use of mechanical and thermal stimulation.

In clinical practice, we know that pain due to a fractured limb could be a challenge because it could be an early step in chronic pain and complex regional pain syndrome type I.21–23We decided to adapt this model from rats to mice because of the accessibility of genetically modified mice. A similar modification of a rat model has been reported by Pogatzki and Raja3to study mechanisms involved in postincisional pain in mice. This approach further allowed these authors to investigate neurobiologic mechanism of pain after surgery, pointing out the interest of using genetically modified mice.24,25 

Our model will probably be useful to study the different pain mechanisms involved in postfracture pain. It will also be valuable to study a variety of pharmacologic treatments. Indeed, as shown in the current study, the pain after bone trauma is sensitive to morphine and NSAIDs. Presumably, peripheral opioid receptors are involved in pain, but the precise mechanism has not been documented yet. It has been reported that opioids resulted in clear analgesic as well as remodeling properties (apoptosis, cell proliferation and growth) in animal models of cancer pain.26Because of the lack of a postfracture pain model, this effect remains unknown in bone trauma and needs to be investigated. The antiinflammatory and antinociceptive activities of NSAIDs are attributed to the inhibition of cyclooxygenase enzymes, decreasing in turn the synthesis of prostaglandins that promote inflammatory responses and enhanced sensitivity to pain at the peripheral site of tissue injury.27It has been shown that NSAIDs impaired fracture healing.28,29Interestingly, the current study indicates that drugs activating opioid receptors and affecting the cyclooxygenase pathway reduce pain behavior in this model. Whether other different systems are also involved will be investigated using different genetically modified strains of mice. Moreover, in these future studies, the current model will probably allow discrimination between the analgesic effect and the remodeling effect of a given drug. However, the limitations of animal models of bone pain analgesics like NSAIDs may impair healing, but with pain relief, weight bearing and increased activity may impair fracture healing as well. Any sustained analgesic effect may affect activity and weight bearing, and this may contribute to changes in remodeling, healing, and so on.

In summary, our model seems to be a reliable and reproducible tool to investigate the effect of closed bone fracture on several parameters, such as pain, remodeling, and recovery. Moreover, it allows studying the effects of various pharmacologic treatments.