Comparative study of a rabbit model of spinal tuberculosis using different concentrations of Mycobacterium tuberculosis

Abstract

BACKGROUND

Tuberculosis is among the most devastating infectious diseases worldwide. Spinal tuberculosis is not easy to detect at an early stage, which without effective treatment often leads to spinal deformity and spinal cord damage which in turn cause complications such as paraplegia and quadriplegia. In this study, we established a model using three concentrations of bacteria and carried out a comprehensive evaluation of the model by imaging, general observations, and histopathological and bacteriological studies.

AIM

To establish a rabbit model of spinal tuberculosis and examine the effect on the model’s efficacy using different concentrations of Mycobacterium tuberculosis (M. tuberculosis) inoculum.

METHODS

New Zealand rabbits were randomly divided into experimental, control and blank groups. The experimental and control animals were sensitized with complete Freund′s adjuvant, a hole was drilled beneath the upper endplate of the L6 vertebral body and filled with gelfoam sponge. The experimental group was divided into three subgroups (experimental 1, experimental 2, experimental 3) and infused with M. tuberculosis suspension at various concentrations. The control group was inoculated with saline and the blank group received no treatment. The 12-week post-operative survival rates were 100%, 80% and 30% in the experimental groups inoculated with concentrations of 10⁶, 10⁷ and 10⁸ CFU/mL bacteria, respectively.

RESULTS

The survival rate of the control and blank groups was 100%. Vertebral body destruction at 8 weeks in the three experimental groups as determined by X-ray analysis was 33.3%, 62.5% and 66.7%, and by computed tomography (CT) and 3-dimensional CT 44.4%, 75% and 100%, respectively. At 12 weeks, the figures were 44.4%, 75% and 100% by X-ray analysis and 44.4%, 100% and 100% by CT and 3-dimensional CT, respectively. All surviving rabbits of the experimental groups had vertebral destruction. The positive bacterial culture rates were 22.2%, 75% and 66.7%, respectively, in the experimental groups. After being sensitized with complete Freund's adjuvant, large differences were observed in the extent of spinal tuberculosis after inoculation of the rabbits with different concentrations of H37RV standard M. tuberculosis.

CONCLUSION

The experimental 1 had a low success rate at establishing an infection. The experimental 3 resulted in high mortality and complication rates. The experimental 2 was optimum for establishing a spinal tuberculosis model based on the high level of symptoms observed and the low rabbit mortality.

Key Words: Spinal tuberculosis; Animal model; H37RV; Mycobacterium tuberculosis; New Zealand rabbits

Core Tip: One study reported having established osteoarticular tuberculosis using non-sensitized animals; however, the result of the experiment was unsatisfactory because of the high mortality rate. In recent studies, Liu and Geng successfully established a spinal tuberculosis model in New Zealand rabbits after sensitizing the rabbits with complete Freund’s adjuvant. However, they established the model without making a comparative study of the effect of using different bacterial concentrations. In this study, we established a model using three concentrations of bacteria and carried out a comprehensive evaluation of the model by imaging, general observations, and histopathological and bacteriological studies.

INTRODUCTION

Tuberculosis is among the most devastating infectious diseases worldwide. Tuberculosis still accounts for the highest mortality from any infectious diseases worldwide, even surpassing human immunodeficiency virus/acquired immunodeficiency syndrome, causing 1.5 million deaths in 2020[1]. According to WHO's report, the current pace of change is not enough: The global cumulative rate of reduction for tuberculosis incidence was only 6.3% between 2015 and 2018, which is much less than that of the 2020 milestone of 20%[2]. Osteoarticular tuberculosis represents 5% of tuberculous infections and of these 50%-60% are spinal tuberculosis, and spinal tuberculosis is the most common subtype of extra-pulmonary tuberculosis, accounting for 15% of cases[3-5]. Spinal tuberculosis is not easy to detect at an early stage, which without effective treatment often leads to spinal deformity and spinal cord damage which in turn cause complications such as paraplegia and quadriplegia[6,7]. The emergence of multi-resistant bacteria has led to the mortality rate due to spinal tuberculosis, as many as 490000 patients have been diagnosed as being affected by multidrug-resistant tuberculosis in 2016 and an additional 110000 tuberculosis people were identified as being resistant to rifampicin, the emblematic most effective first-line antitubercular drug so far[8]. Therefore, it is essential to establish animal models of osteoarticular tuberculosis successfully to treat the disease and conduct future research; however, thus far few successful models have been reported[9,10].

Choosing a suitable animal model to simulate the development of human disease is extremely important, needing to be suitable for studies involving inter alia pathogenesis, pathology, immune mechanisms and drug therapy. Local injection of tuberculosis bacteria without using anti-tuberculosis drugs often causes a fatal bacteremia. However, the stringent laboratory conditions required when working with Mycobacterium tuberculosis (M. tuberculosis) greatly hinder progress.

One study reported having established osteoarticular tuberculosis using non-sensitized animals; however, the result of the experiment was unsatisfactory because of the high mortality rate. In recent studies, Liu et al[9] and Geng et al[11] successfully established a spinal tuberculosis model in New Zealand rabbits after sensitizing the rabbits with complete Freund’s adjuvant. However, they established the model without making a comparative study of the effect of using different bacterial concentrations. In this study, we established a model using three concentrations of bacteria and carried out a comprehensive evaluation of the model by imaging, general observations, and histopathological and bacteriological studies.

MATERIALS AND METHODS

Experimental conditions and equipment

The study fulfilled the requirements of the 309^th Hospital of Chinese PLA (Army Tuberculosis Prevention and Control Key Laboratory, Beijing Key Laboratory of New Techniques or Tuberculosis Diagnosis and Treatment, Institute for Tuberculosis Research). The principle of aseptic surgical techniques was strictly followed throughout the experiment. The experimental protocols and animal welfare were approved by the Animal Ethics Committee at the 940^th Hospital of Joint Logistic Support Force of Chinese People’s Liberation Army.

M. tuberculosis strain H37RV was obtained from the Army Tuberculosis Prevention and Control Key Laboratory of the 309th Hospital of Chinese PLA. Instruments utilized included a Computed Radiography Digital Imaging System (Philips, Eindhoven, Netherlands) and a 16-slice computed tomography (CT) scanner (United Imaging, Shanghai, China). The main reagents included complete Freund’s adjuvant (Sigma-Aldrich, St. Louis, MO), Lowenstein-Jensen (LJ) medium and Petri dishes (Baso, Zhuhai, China).

Experimental animals and grouping

Forty-two healthy New Zealand white rabbits (aged 11-12 weeks, of mixed sex, with body weights of 2.25 ± 0.15 kg) were obtained from the Center for Animal Experiment of the 309th Hospital of Chinese PLA. The license number for medical experimental animal production was SYXK(Army)2012-0016. The rabbits were randomly divided into five groups: Three experimental groups of 10 each The experimental group was divided into three subgroups (experimental 1, experimental 2, experimental 3, inoculated with concentrations of 10⁶, 10⁷ and 10⁸ CFU/mL bacteria, respectively), 10 in the control group and two in the blank group. During the experiments, all the rabbits were raised separately in individual cages with a standardized living environment and feeding pattern. At the end of the experiment, the rabbits were killed (pentobarbital sodium was injected into the auricular vein at a dose of 100 mg/Kg), sealed and cremated. Tuberculin-purified protein derivative 5IU tests were carried out before the experiment and only animals with negative results were used for the study.

Mycobacterial suspensions

H37RV standard M. tuberculosis was cultured for 4 weeks in LJ medium. Well-established colonies were selected and triturated in 0.9% saline. Bacterial suspensions were collected and diluted with saline to obtain suspensions of 10⁶, 10⁷ and 10⁸ CFU/mL as determined by the colony counting method. These were maintained in a 4 °C refrigerator (none for more than 24 hours) and shaken well before application.

Preparation of an animal model

Sensitization with complete Freund’s adjuvant: The experimental and control rabbits received 0.1 mL of complete Freund’s adjuvant (a mixture of lanolin and liquid paraffin containing 4.5 mg/mL bovine M. tuberculosis) through an intradermal injection in the nape of the lower neck. Two animals developed anorexia, fever, spirits are drooping, activities were significantly reduced and diarrhea 2 weeks after injection and were given intramuscular injections of gentamicin for 5 days without significant improvement. Finally, the two animals were removed from the experiment and replaced (air embolism was applied after anesthesia in the rabbits to minimize pain and distress). Rabbits in the blank group received no treatment. We performed the experiments 1 month later.

Anesthesia: Pentobarbital sodium was injected into the auricular vein at a dose of 30-40 mg/Kg. None of the animals showed obvious discomfort during surgery.

Operation: The rabbits were given small amounts of water and fasted for 12 hours pre-operatively. After the anesthesia had taken effect, the rabbit was placed on an operating table and the right lower back was shaved, disinfected twice with Povidone-iodine and covered with aseptic towel. A longitudinal incision of 5-6 cm was made along the end of the left 12^th rib inferior to the iliac crest, the subcutaneous fascia and muscle were separated carefully to avoid injury to vessels and nerves, and the intervertebral disc of L5/6 and adjacent vertebral bodies was completely exposed using the method of Geng et al[11]. An electric drill was used to bore a hole (3-mm diameter, 6-mm depth) from the right front to the left rear of the L6 vertebral body at a 30° angle to the transverse plane. After hemostasis, the cavity was filled with gelfoam sponge, then 0.1-mL aliquots at 10⁶, 10⁷ or 10⁸ CFU/mL were slowly infused into the gelfoam sponge, using 10 rabbits for each of the three bacterial concentrations. The control group each received 0.1 mL 0.9% saline, while no treatment was performed on the blank group. Finally, the incisions were sutured layer by layer, and sterile drape was fixed over the incision.

After surgery, the rabbits were housed separately following the principles of aseptic technique, safety and animal ethics.

Study parameters and methods

General observations: The animals were observed twice daily for state of health, eating habits, mental state, activity and wound healing.

Imaging examinations: X-ray analysis, CT and 3-dimensional CT (3DCT) were performed at 8 and 12 weeks post-operatively (po) to observe the extent of destruction of the intervertebral disk and vertebral body and the formation of sequestra and abscesses. All these imaging examinations were performed under anesthesia.

Histopathological observation: Rabbits were sacrificed 12 weeks po after imaging examinations had been completed. We dissected out lesion tissue (including the intervertebral disk of L5/6, the upper and lower endplates, the lesions' vertebral body and soft tissue, and abscesses) for routine hematoxylin and eosin (HE) staining to detect any histopathological changes. Soft tissue and bone tissue were fixed in 10% buffered formalin for 3-5 days before routine processing. After the fixation had been completed, bone tissue decalcification was performed with 10% nitric acid, followed by dehydration, embedding, sectioning, staining with HE and microscopic observation.

Culturing of M. tuberculosis: Diseased tissue (0.5 g of granulous tissue or paravertebral abscess) was removed from the operation site of all surviving rabbits in the experimental and the control group, cut into pieces and put into 1 mL 0.9% saline. After homogenization, 1 mL of 6% sodium hydroxide (NaOH) was added to the mixture and allowed to stand at room temperature for 30 minutes to kill the non-tuberculous mycobacteria. Finally, a further 1 mL 0.9% saline was added to the mixture, from which 0.1 mL was collected and cultured on LJ medium at 37 °C for 4-8 weeks. The results were analyzed according to the standards for tuberculosis diagnosis and bacteriological testing published in 1995 by the Chinese Tuberculosis Prevention and Treatment Association[12].

Statistical analyses

All results were expressed as the mean ± SEM. The Fisher exact probability method was used for the 12-week post-operative animal survival rate.

RESULTS

General observations

Experimental 1: Of the 10 rabbits inoculated, one rabbit suddenly appeared opisthotonus and died of cardiopulmonary arrest when we drilled the hole, possibly due to spinal cord or nerve injury. The other rabbits underwent surgical procedures smoothly, two of them eating poorly and appearing subdued three days after surgery. Thereafter, the general condition of the rabbits gradually improved and no infection was observed at the incision. Finally, nine rabbits survived to the end of the experiment, three of them showing weight loss, anorexia and reduced activity. The remainder showed no obvious signs of disease.

Experimental 2: Of the 10 rabbits, one became paraplegic the next day and spirits are drooping, activities were significantly reduced, so air embolism was applied after anesthesia to minimize pain and distress po at 7 days. Three rabbits ate poorly and had reduced activity, one of which weight significantly reduced, spirits are drooping, activities were significantly reduced and the fur ruffled, so air embolism was applied after anesthesia po at 35 days. On dissection, the dead rabbit had pus within the chest and nodules of varying sizes on the lung surface. The other body parts showed no obvious abnormalities. Finally, eight rabbits survived to the end of the experiment, all experiencing weight loss, anorexia and reduced activity.

Experimental 3: Of the 10 rabbits, seven experienced mental fatigue and ate poorly, four of them having nonunions and pus at the incision site and dying 12-20 days po. Dissection of the dead bodies yielded huge amounts of pus in the surgical site and right lung with caseous necrosis (Figure 1). Three rabbits each lost about 1 kg of weight after about 1 month and died 38-51 days po (one of which were euthanized with air embolism after anesthesia, the others died a natural death). Dissection of the rabbits yielded serious swelling at the surgical site, right lumbar region and the right thigh. Off-white pus from the lumbar region was cultured on LJ medium at 37 °C for 4 weeks and good colony growth was observed. The cause of death was ascribed to disseminated M. tuberculosis. Finally, three rabbits survived to the end of the experiment, all three showing weight loss, anorexia and reduced activity.

Figure 1 Abnormal lung with white caseous lesions, serious lung tissue damage and abnormal anatomy. The contralateral lung in ruddy color had normal anatomical structures with no obvious pathological changes.

Control group: Of the 10 rabbits, one died because of an anesthetic accident, while nine survived to the end of the experiment, none of them showing weight loss, anorexia or reduced activity.

Blank group: The animals maintained a good appetite with normal activity and zero mortalities. Animal survival rates are shown in Table 1.

Table 1 Animal Survival in the five groups.

Groups (n)	Day of model establishment	4 weeks post model establishment	8 weeks post model establishment	12 weeks post model establishment	Survival rate (%)
Experimental 1 (n = 10)	9	9	9	9	100
experimental 2 (n = 10)	10	9	8	8	80
experimental 3 (n = 10)	10	6	3	3	30
Control group (n = 10)	9	9	9	9	100
Blank group (n = 2)	2	2	2	2	100

The Fisher exact probability method was used for the 12-week post-operative animal survival rate. The experimental groups (1 and 2) compared with the control group: P > 0.05. The experimental group 3 compared with the control group: P < 0.05. The experimental group 1 compared with the experimental group 2: P > 0.05. The experimental group 1 compared with the experimental group 3: P < 0.05. The experimental group 2 compared with the experimental group 3: P < 0.05.

Imaging results: Typical results of X-ray analysis, CT and 3DCT scans performed at 12 weeks are shown in Figure 2. The number of animals with bone destruction is recorded in Table 2. X-ray examinations revealed that, in the experimental groups, the intervertebral spaces appeared blurred and narrowed, while the adjacent endplates appeared to have increased bone density shadows (Figure 2A and B). CT scans showed that the bone mineral densities (BMD) in the vertebral bodies were not homogeneous and that some areas had irregularly low BMDs. Dotted sequestra and increased BMD shadows were also visible in the vertebral bodies, while the local vertebral bone cortices were nonunion. The soft tissue did not show obvious swelling around the vertebral bodies (Figure 2C-E). 3DCT showed that the intervertebral spaces had narrowed with surrounding osteophyte formation (Figure 2F). X-ray analysis of the control group showed no post-operative changes in the intervertebral spaces and no damage to the vertebral bodies at either 8 or 12 weeks. CT scans revealed that the vertebral body edges were slightly hardened with no evidence of bone destruction; the intervertebral spaces had not changed and there were no abscesses or soft tissue calcification shadows around the vertebral bodies. There was no abnormal imaging observed in the blank group.

Figure 2 Twelve-week post-operative imaging results. A and B: X-ray analysis revealed that the intervertebral spaces appeared blurred and narrowed, the adjacent endplates appeared to have increased bone density shadows (orange arrows); C-E: Computed tomography (CT) scans showed that dotted sequestra and increased bone mineral densities shadows could be observed in the vertebral bodies; the local vertebral bone cortices were nonunion (orange arrows). There was no obvious soft tissue swelling around the vertebral bodies; F: 3-dimensional CT scans showed that the intervertebral spaces had become narrowed with surrounding osteophyte formation (orange arrows).

Table 2 Positive post-operative image changes in the three groups of experimental animals at 8 and 12 weeks.

Results	Experimental 1		Experimental 2		Experimental 3
	8 weeks	12 weeks	8 weeks	12 weeks	8 weeks	12 weeks
X-ray
Intervertebral space changes	3	4	5	6	2	3
Osteophyte	1	2	4	4	2	3
CT
Osteophyte	2	3	5	8	3	3
Destruction of the vertebral body	4	4	6	8	3	3
Abscess	0	0	0	0	0	0
3-dimensional CT
Intervertebral space changes	4	4	6	8	3	3
Osteophyte	2	3	5	8	3	3

CT: Computed tomography.

General observations: We dissected all animals 12 weeks po after completing the imaging examinations.

Experimental 1: The incisions of all the animals had healed well, with no fistulas or discharge phenomena being observed. Four rabbits with positive imaging results were found to have different extents of vertebral body destruction, but did not show significant sequestrum formation. Intervertebral disc color had darkened and intervertebral space had narrowed; soft tissue around the vertebral bodies had been destroyed to varying degrees accompanied by a little light yellow pus, but no obvious psoas abscesses occurred in these rabbits. Caseous necrotic substances and granulous tissue were not found within the paravertebral tissue.

Experimental 2: One rabbit’s incision had not healed well, showing a little light yellow pus. Obvious bone destruction had occurred on the inferior border of the L5 vertebra or the superior border of the L6 vertebra in all animals; the intervertebral spaces had become narrowed with obvious destruction of the intervertebral discs. Two rabbits were found with sequestra within the vertebral bodies, with the vertebral sides having scattered small nodules with caseous necrotic material overflowing after piercing. No obvious psoas abscesses occurred in these rabbits.

Experimental 3: The incisions of two of the rabbits had not healed well and fistulas with diameters of about 0.2 cm were found near the incision, caseous necrotic material emerging when squeezed. Serious vertebral body destruction had occurred and the intervertebral spaces had narrowed; the intervertebral discs were extensively damaged and had become dark. Scattered small nodules with caseous necrotic material that overflowed on piercing were present on both sides of the badly damaged paravertebral muscle. No obvious psoas abscesses occurred in these rabbits. Two rabbits had slightly damaged vertebral bodies and intervertebral discs among the five animals with negative imaging results. No other organs with obvious abnormalities were observed in any of the animals. No obvious destruction of the intervertebral disks or vertebral bodies was detected and no abscesses were found in either the control group of rabbits, the blank group or those rabbits with negative imaging results.

Histopathological observations: After the experiment is completed , air embolism was applied after anesthesia in the rabbits to minimize pain and distress .Histopathological studies of the rabbits with positive imaging results showed pathological changes typical of tuberculosis, with vast areas of caseous necrosis (Figure 3A) and inflammatory cell infiltration (Figure 3B) in the affected vertebral bodies. There were multiple sequestra and the structure of trabecular bone was either disordered or had disappeared (Figure 3C). Typical tuberculous nodules and caseous necrotic material were observed in the center surrounded by epithelioid cells, multinucleated giant cells and many infiltrated inflammatory cells (Figure 3D). Many epithelioid cells and typical multinucleated giant cells surrounding the caseous necrotic material were visible (Figure 3E). In the control group, the trabecular bone structure was normal with no sequestra or epithelioid cell formation being observed (Figure 3F).

Figure 3 Twelve-week post-operative histopathological results using hematoxylin and eosin stain. A: Pathological changes typical of tuberculosis with vast areas of caseous necrosis [diamond; hematoxylin and eosin (HE) × 200]; B: Infiltration of inflammatory cells (triangle) and caseous necrosis (diamond; HE × 200); C: Disordered trabecular bone structure and numerous sequestra (orange arrow; HE × 100); D: Typical tuberculous nodules with caseous necrotic material in the center: Epithelioid cells (orange arrow), multinucleated giant cells (black arrow) and inflammatory cell infiltration (triangle; HE × 400); E: Typical multinucleated giant cells (black arrow) and epithelioid cells (orange arrow) were observed (HE × 400); F: No abnormalities were observed in the control group (HE × 200).

M. tuberculosis culture: After the experiment is completed , air embolism was applied after anesthesia in the rabbits to minimize pain and distress. Paravertebral pus, soft tissue and necrotic tissue were taken from all experimental animals for culture of M. tuberculosis. After culturing for 4 weeks, 10 of the 20 rabbits in the experimental groups had positive results with the appearance of pale yellow colonies (two of the experimental 1, six of the experimental 2 and two of the experimental 3). Yellow colonies were evenly and firmly attached to the medium (Figure 4). No mycobacterial growth was found in the other 10 specimens after 8 weeks of culture. The rates of positive culture were 22.2%, 75% and 66.7% for the above three groups, respectively. No mycobacterial growth was observed with either the control or blank groups.

Figure 4  Pale yellow granular colonies were evenly and firmly attached to the Lowenstein-Jensen medium.

DISCUSSION

Choosing a suitable animal is critical to the success of establishing a model for spinal tuberculosis. Many animals have been tried for establishing a model, including mice, guinea pigs, rabbits, monkeys, cows, sheep and fish[13-15]. As early as 1928, Lurie successfully established the first pulmonary tuberculosis model[16]. Rabbits are sensitive to M. tuberculosis and can form typical tuberculous granulomas after infection[17] and are commonly used to establish models for pulmonary tuberculosis[18]. Additionally, the rabbit spine is similar to that of humans, in that the relatively bulky lumbar vertebrae each consist of an anterior vertebral body, rear vertebral plates, and spinous and ambilateral transverse processes. These features make the rabbit spine suitable for establishing a spinal tuberculosis model, on which it is possible to perform spinal surgery, debridement and bone graft operations. In this study, we successfully used the New Zealand rabbit to establish a model of spinal tuberculosis and to determine the optimum bacterial concentration to achieve this end.

Local direct inoculation with M. tuberculosis results in different pathological outcomes post-infection. If the bacterial concentration in the inoculum is too low, the success rate of the model, and of obtaining positive images, is low and the bone tissue mainly presents as hyperplasia. However, large concentrations of bacteria rapidly cause blood bacteremia and caseous pneumonia, which lead to early mortality and a low success rate. Therefore, the appropriate quantity of bacteria, the necessary prior interventions and the correct infection pathway are the keys to establishing a successful model.

H37RV standard M. tuberculosis is widely used in tuberculosis research[19-21]. We kept the bacteria at 4 °C in a refrigerator to guarantee viability. Three concentrations of bacteria were used in this study. The survival rate of rabbits in the experimental group inoculated with 10⁶ CFU/mL was 90% and positive po image changes totaled 44.4% by 12 weeks. The survival rate in the group inoculated with 10⁷ CFU/mL was 80% and a successful model was established po in all the surviving rabbits by 12 weeks. The survival rate in the group inoculated with 10⁸ CFU/mL was only 30% and successful model establishment was achieved post-operatively in the surviving rabbits by 12 weeks. In conclusion, a M. tuberculosis concentration of 10⁷ CFU/mL was suitable for establishing a spinal tuberculosis model in the rabbit with a higher survival and success rate.

Complete Freund’s adjuvant is an oily emulsion enriched with bovine heat-killed M. tuberculosis. The slow release and continuous stimulation by immunogenic substances enhances the systemic resistance of the host to M. tuberculosis. Moreover, local lesions on the vertebral structure are aggravated by type IV immune hypersensitivity reactions, which enhance bone destruction. Some researchers have used unsensitized animals to establish models of osteoarticular tuberculosis, but because the local and systemic responses were more severe, rapid mortality of the experimental animals occurred. Wu and Duan[22] and Li[23] sensitized rabbits using complete Freund’s adjuvant to establish a model of knee tuberculosis and got satisfactory experimental results because of the lighter systemic and local reactions. Likewise, when we established the model for osteoarticular tuberculosis, the use of complete Freund’s adjuvant significantly increased the animals’ survival and the experimental success rate, making it worthy of becoming a generalized procedure.

As early as 1928, John Fraser had tried local, venous and arterial implanting of M. tuberculosis to establish a model of osteoarticular tuberculosis, but the survival of, and success rate with, the animals was very low. Subsequently, many scholars have used the method of directly implanting M. tuberculosis in the nutrient foramen. A hole was drilled in the long bone for local implanting in a sensitized animal, which procedure successfully established osteoarticular tuberculosis models, but the success rate was not very satisfactory. Wu et al[22] successfully established a rabbit model of knee joint tuberculosis for the first time in China by injection of 0.5 mg bovine mycobacterium in 0.1 mL. Li[23] established a rabbit model of knee joint synovial tuberculosis by implanting H37RV standard M. tuberculosis using a similar method. Liu et al[24] successfully established a spinal tuberculosis model by implanting H37RV standard M. tuberculosis (10⁷ CFU/mL in 0.1 mL), the survival and success rate both being satisfactory. Shang et al[25] established a spinal tuberculosis model successfully with a similar method, implanting rifampin-chitosan-calcium alginate sustained-release microspheres into the animals. This produced a better anti-tuberculosis response. In this study, we established a spinal tuberculosis model successfully but with different concentrations of M. tuberculosis and were able to determine a suitable amount of inoculating bacteria.

An early diagnosis of spinal tuberculosis is often difficult because of the lack of specific clinical manifestations. Imaging examination is an important method for evaluating the spinal tuberculosis model and has made earlier diagnosis possible. In our study, X-ray presentations were blurring and narrowing of the intervertebral space, with uneven vertebral density and new bone formation at the margin of the vertebral body; however, we had few positive results and most appeared relatively late. CT imaging is superior to plain radiography for demonstrating the presence of paraspinal abscesses. Uneven bone density, irregular areas of decreased bone density, dotted sequestral shadows, increased bone density within the vertebral body and uneven partial vertebral bone cortices were observed. 3DCT is a superior method for detecting narrowing of the intervertebral spaces and osteophytes.

Histopathological examination showed vast areas of caseous necrosis and inflammatory cell infiltration in the vertebral bodies. There were numerous sequestra within the vertebral bodies and the structure of the trabecular bone was disordered or absent; multinucleated giant cells, typical tuberculous nodules and epithelioid cells also could be observed in some animals.

While imaging and histopathological examination are highly suited for the detection of spinal tuberculosis, mycobacterial culture remains the diagnostic “gold standard”. However, mycobacterial culture cannot be used to obtain an accurate diagnosis in a short-term study, due to the slow growth rate of the bacteria and the low frequency of detection. In our study, paravertebral pus, soft tissue and necrotic tissue were used as a source of bacteria for culture. The rate of positive culture was 22.2%, 75% and 66.7%, respectively, from the groups inoculated with 10⁶, 10⁷ and 10⁸ bacteria per mL.

We used the method of implanting the mycobacterial suspension by infusion into a gelfoam sponge that had been inserted into the vertebral body, using three concentrations of M. tuberculosis in separate experimental groups. Finally, we concluded that, of the three concentrations used, injection of 0.1 ml at 10⁷ CFU/mL was superior for establishing a spinal tuberculosis model, because of the high rate of disease signs accompanied by low rabbit mortality. In conclusion, by inoculating local vertebral bodies of New Zealand white rabbits with an appropriate dose of strain H37RV, we established a model for spinal tuberculosis, the pathological changes of which were similar to those observed in human spinal tuberculosis. This method was simple with a relatively high success rate, potentially providing the means to facilitate studies on the pathology of spinal tuberculosis and on the search for new drugs or a vaccine.

The limitations of our study included the small sample size for each group (laboratory facilities were limited) and the large M. tuberculosis concentration range employed; thus, we cannot say with certainty that 0.1 mL of bacteria at 10⁷ CFU/mL was the most suitable for establishing the spinal tuberculosis model. Therefore, further studies should be carried out by improving the experimental conditions, employing larger numbers of test animals, and doing a comparative study with a smaller concentration gradient of bacteria, to determine an accurate figure for the optimum bacterial dose for the establishment of a rabbit model of spinal tuberculosis.

CONCLUSION

The experimental 1 had a low success rate at establishing an infection. The experimental 3 resulted in high mortality and complication rates. The experimental 2 was optimum for establishing a spinal tuberculosis model based on the high level of symptoms observed and the low rabbit mortality.