A novel rat model of vertebral inflammation–induced intervertebral disc degeneration mediated by activating cGAS/STING molecular pathway

Abstract In this study, we describe a new rat model of vertebral inflammation–induced caudal intervertebral disc degeneration (VI‐IVDD), in which IVD structure was not damaged and controllable segment and speed degeneration was achieved. VI‐IVDD model was obtained by placing lipopolysaccharide (LPS) in the caudal vertebral bodies of rats. Rat experimental groups were set as follows: normal control group, group with a hole drilled in the middle of vertebral body and not filled with LPS (Blank group), group with a hole drilled in the middle of vertebral body and filled with LPS (Mid group), and group with hole drilled in the vertebral body in proximity of IVD and filled with LPS (NIVD group). Radiological results of VI‐IVDD rats showed a significant reduction in the intervertebral space height and decrease in MRI T2 signal intensity. Histological stainings also revealed that the more the nucleus pulposus and endplate degenerated, the more the annulus fibrosus structure appeared disorganized. Immunohistochemistry analysis demonstrated that the expression of Aggrecan and collagen‐II decreased, whereas that of MMP‐3 increased in Mid and NIVD groups. Abundant local production of pro‐inflammatory cytokines was detected together with increased infiltration of M1 macrophages in Mid and NIVD groups. Apoptosis ratio remarkably enhanced in Mid and NIVD groups. Interestingly, we found a strong activation of the cyclic GMP‐AMP synthase /stimulator of interferon gene signalling pathway, which is strictly related to inflammatory and degenerative diseases. In this study, we generated a new, reliable and reproducible IVDD rat model, in which controllable segment and speed degeneration was achieved.


| INTRODUC TI ON
About 18%-48% of the population will experience chronic low back pain (LBP) at some time during life, and most LBP patients show intervertebral disc degeneration (IVDD). 1,2 IVDD is a complex age-related process 3 with a multifaceted aetiology including signs such as inflammation, 4 micro-damage, 5 biomechanical stress, 6 ageing, 7 apoptosis 8 and autophagy. 9 However, the integrated molecular mechanisms underlying IVDD have not been fully elucidated, 10 and no effective and reliable intervertebral disc (IVD) repair strategy has been developed so far. 11 As a result, symptomatic or alternative treatments are often adopted for IVDD-induced spinal diseases in clinic. Hence, much efforts are needed to identify the mechanisms underlying IVDD and potential therapeutical targets, and the generation of highly reliable and reproducible animal models of IVDD will be fundamental.
Due to factors, such as animal ethics, source, costs, human similarity and ease of manipulation, the most commonly used animals for establishing IVDD models are rats/mice and rabbits. 12 IVDD animal models are mainly divided into two categories: induced IVDD model (abnormal stress model, 13 spinal instability model, 14 mechanical damage model, 15 chemical induction model 16 and biological induction model 17 ) and spontaneous IVDD model (endocrine change model 18 and standing model 19 ).
Attempts have been made to induce IVDD by changing the IVD stress in many studies. Lai et al. 13 induced caudal IVDD in rats through abnormal axial stress caused by pressurizing the rat tail. Our previous studies showed that IVDD could also be obtained by applying shear force to rabbit IVDs. 20 Meanwhile, Miyamoto et al. 14 generated IVDD by surgically inducing the instability of the spine. The degeneration model dependent by stress change strongly resembles IVDD, since it simulates the effects of stress change caused by different postures. The standing model of spontaneous IVDD is similar to these models. Ao et al. 19 developed an animal model in which IVDD was induced by continuously stimulating the mice to take bipedal standing posture, perfectly simulating the process of natural IVDD induced by human standing. However, the long-time duration [(1 h/day) × 1 month to (6 h/day) × 2 months) and low success rate limit research on IVDD condition in these models.
To simulate the process of natural IVDD in humans, gene editing or ovariectomy 17,18 has also been used to induce IVDD in animals. However, the high costs, as well as the scarce experimental reproducibility and reliability, represent a limit of these methods.
Hence, there is a need to establish easy-to-handle models, with low costs and high success rate, such as those induced by chemicals and mechanical damage. Chemical induction models are obtained by injection of protease, collagenase or bleomycin into the IVD. 16,21 The most typical mechanical damage model is the rat caudal IVD acupuncture model, which is currently also the most widely applied model. 8,15 However, irreversible damage to IVD structure is main limitation of these models. In fact, according to previous and our observations, 8,22 once the IVD structure is damaged, degeneration rapidly and irreversibly progresses. Animal models in which IVD structure is compromised cannot simulate natural IVDD human condition, such as age-related IVDD.
A large number of studies 8,15,23 have shown that IVDD models induced by mechanical damage are highly reproducible and dependent on local inflammatory responses induced by IVD damage. Although inflammation is an important factor for rapid induction of IVDD, 4 IVD structure damage obtained by mechanical stress in animals is unsuitable for studying natural degenerative process.
In this study, we describe a new, highly reliable and reproducible IVDD rat model obtained by inducing inflammation in the vertebral body without affecting the integrity of IVD structure. In particular, IVDD was induced by placing the pro-inflammatory agent lipopolysaccharide (LPS) in the middle of the caudal vertebral bodies of rats. LPS treatment determined the triggering of rapid, stable and localized inflammatory responses, 24 with concomitant degeneration of the adjacent IVD. In this vertebral inflammation-induced IVDD (VI-IVDD) model (Figure 1), the progression of inflammation was studied by using radiological imaging, as well as by performing morphological, histological and cytological evaluations. The involvement of LPS-induced cyclic GMP-AMP synthase (cGAS)/stimulator of interferon genes (STING) signalling pathway 25 was also investigated in the model.

| Animals
Ten to 12-week-old male SPF-grade Sprague Dawley rats (Shanghai SLAC Laboratory Animal Co., Ltd.; n = 192), with an average bodyweight of 250-300 g, were randomly divided into four groups (n = 48 per group): normal control group (Normal group), group with a hole drilled in the middle of vertebral body and not filled with LPS (Blank group), group with a hole drilled in the middle of vertebral body and filled with LPS (Mid group, also called VI-IVDD basic model), and group with hole drilled in the vertebral body in the proximity of IVD and filled with LPS (in the proximity of intervertebral disc, NIVD group, also called VI-IVDD acceleration model). The progression of inflammation and degeneration were studied at 1, 2 and 4 weeks after surgical intervention, with 16 rats being randomly selected for analysis for each group and time-point.
No accidental death or operation failure occurred in rats used for the study. Skin surgical incision in rats healed well, and no infections were observed at cutaneous or subcutaneous soft tissue levels.
At the end of experimental procedure, rats were anaesthetized and killed by cervical dislocation.
All surgical interventions, treatments and animal care procedures were performed in strict accordance with the Animal Care and Use Committee of the University School of Medicine.
Caudal vertebral bodies, IVDs and large vessels have been previously localized in the caudal region of rats ( Figure 1A). In particular, the 9th or 10th caudal vertebral bodies were localized by touching the ilium and sacrum of the rats, and the area where the hair disappeared at the junction of trunk and tail (in proximity of the 9th caudal vertebral body). Four large subcutaneous vessels in the tail of rats, visible to naked eyes, have also been localized in the dorsal and ventral midlines, and in both sides from the middle of the tail.
Normal group rats were not treated, whereas the other three groups were subjected to hole drilling into vertebral bodies ( Figure 1B). For drilling, it was necessary to make a longitudinal incision (0.5 cm) in the skin, carefully avoiding vessels. The subcutaneous fascia and ligament were separated, and the modelling vertebral body was exposed ( Figure 1CI,II). In Blank and Mid groups, a drill bit (diameter: 1.5 mm) was drilled into the middle of the vertebral body, with an inclination of 45°, without going through the vertebral body.
In the NIVD group, the hole has been drilled into the vertebral body in the proximity of one side of IVD (about 2-4 mm; Figure 1B,C-III). After drilling, 0.25 mg of LPS powder (Escherichia coli O111:B4, Sigma-Aldrich Co., Ltd.) was placed into the drilled hole in rats of Mid and NIVD groups and sealed with bone wax ( Figure 1C-IV) in order to prevent LPS leakage out of the vertebral body and infections of the subcutaneous soft tissue. Based on previous team experience, IL-1β, IL-6 and TNFα cytokines were not used as inflammatory triggers in the model, as they are high-cost, easily degradable, not capable to sustain local inflammatory responses and quite often responsible for systemic inflammation. Skin was then sutured and disinfected. When the rats were fully awake, they were moved to the animal room and treated normally. The incision skin was daily disinfected, and the local pain was relieved with a mixture solution F I G U R E 1 Intervertebral disc degenerationmodel (IVDD) establishment. (A) Anatomy of rat tail; (B) experimental groups included the normal control group (Normal group), group with hole drilled in the middle of vertebral body without placing lipopolysaccharide (LPS) (Blank group), group with hole drilled in the middle of vertebral body and treated with LPS (Mid group, also called VI-IVDD basic model), and group with hole drilled at one side (in proximity of IVD) of the vertebral body and treated with LPS (NIVD group, also called VI-IVDD acceleration model); (C) experimental procedure included the following steps: (I) localization of vertebral body, IVD and vessel; (II) dissection and localization of subcutaneous tissue and caudal ligament; (III) identification of drilling position, and execution of drilled holes in vertebral bodies with a drill bit (diameter: 1.5 mm), inserted with a 45-degree inclination; and (IV) in Mid and NIVD groups, LPS was placed and sealed with bone wax, whereas in the Blank group, drilled hole was sealed with bone wax without LPS. After surgical intervention, skin was sutured and disinfected of 75% alcohol and 20 g/ml lidocaine (1:1) to prevent skin infections and impact on experiments.

| Radiological evaluation
Caudal IVD and intervertebral space of rats of different groups were radiologically evaluated at the different time-points. According to our previous research, molybdenum-target plain X-ray image is superior to plain X-ray in assessing intervertebral space in rats. 22 Hence, molybdenum-target plain X-ray system (Mammomat Inspiration, Siemens) was adopted to obtain images of vertebral bodies and intervertebral space. Disc height index (DHI) was measured to evaluate the change in intervertebral space height and calculated as follows: Figure S1). 8,22,27 Magnetic resonance imaging (MRI) (uMR770; Shanghai United Imaging Intelligence Healthcare Co., Ltd.) was performed on the tail of rats, and Adobe Photoshop software (CC 2018, Adobe Systems Incorporated) was used to measure the total greyscale value of IVD adjacent to the operating vertebral body (a smaller value indicated severer degeneration), 28 and thus to quantitatively analyse the severity of IVDD.

| Sampling
The operated vertebral body, the adjacent IVD and a half of the adjacent vertebral body were kept and fixed with 4% paraformaldehyde.
Samples were then decalcified, dehydrated and embedded in paraffin. Four-micrometer sections were deparaffinized and preserved on a glass slide at room temperature (pathology slicer and Leica embedder provided by Shanghai Leica Instrument Co., Ltd.) until stainings were performed.

| Immunofluorescence
M1 macrophage markers, such as inducible nitric oxide synthase (iNOS) and CD68, 31,32 cGAS and STING (endoplasmic reticulum protein) proteins of the cGAS/STING signalling pathway, and TANKbinding kinase 1 (TBK1), 25,33 were detected by immunofluorescence (IF). Double-IF labelling was adopted for iNOS, CD68, cGAS and STING. 34 After quenching endogenous peroxidase with 3% H 2 O 2 , achieving antigen retrieval and blocking nonspecific binding sites by serum, sections were incubated at RT with primary (macrophage: iNOS, signalling pathway: cGAS) and then secondary antibodies. in accordance with the above procedures, DAPI counterstaining and subsequent steps were carried out. The antibodies were all provided by ABclonal Co., Ltd. AOD of each marker was measured using the Image-Pro Plus 6.0 software.

| Statistical analysis
All data were analysed using IBM SPSS Statistics for Windows, version 20 (IBM Corp.). p < 0.05 was considered statistically significant. According to data type (obvious interaction between groups and time variables), simple effect analysis of factorial design was adopted to analyse the DHI of molybdenum-target X-ray images (interaction: F = 14.564, p < 0.05; Table 1), the total greyscale value of IVD in MRI (interaction: F = 26.000, p < 0.05; Table 1

| Radiological evaluation
Vertebral body, adjacent intervertebral space and the drilling position were firstly visualized by molybdenum-target plain X-ray system in rats of the four experimental groups (Figure 2A). X-ray images showed that holes were drilled accurately in Blank, Mid and NIVD groups and there was no direct damage to the IVD. Normal and Blank groups had a normal intervertebral space height, and no evident changes from 1 to 4 weeks were observed. Mid and NIVD groups had a significant smaller intervertebral space height as compared to the other two groups, and it substantially decreased over time. Of note, height of intervertebral spaces adjacent to both sides of vertebral body in Mid group seemed to be reduced, whereas height of the intervertebral spaces in the proximity of the drilled hole, but not that distant, shows clear changes in the NIVD group. The quantitative index of intervertebral space height DHI also changed in Mid and NIVD groups, and in particular, it significantly decreased over time, as shown in Figure 2B and Table 2 (p < 0.05, differences in DHI between Mid or NIVD group and Normal or Blank group at 1, 2 and 4 weeks). However, no statistically significant difference was observed between Normal and Blank groups (p > 0.05). Moreover, after 1 week, the decrease in intervertebral space height in the NIVD group was greater than that measured in the Mid group (mean difference = 2.942, p < 0.05, 2 weeks; mean difference = 2.198, p < 0.05, 4 weeks).
As shown in Figure    visible as dark red and light blue, respectively. In Mid and NIVD groups, AF was more disordered, and light blue staining of cartilage EP disappeared. As shown in Figure 4B, blue staining area of the degenerated IVD extended at late time-point of observation, indicating that immature fresh collagen fibres were gradually formed (immature and mature collagen fibres stain blue and red, respectively). At week 4, the newly formed collagen fibres showed maturation signs.

| Macrophage infiltration analysis
We next evaluated M1 macrophage infiltration by analysing iNOS and

| Progression of inflammation and degeneration
In can be realized with a high success rate and at low cost, and more importantly is easily operable, so that it is one of the most favourite animal model in many studies on IVDD. However, in this model, the IVD structure is highly damaged, determining a degeneration progress too rapid and irreversible due to strong inflammatory responses in the damaged site. 8,15,23 The damage model cannot be employed in studies aimed at identifying mechanisms involved in age-related natural IVDD.
A large body of literature supports that inflammation is a major inducer and accelerator of IVDD pathological processes. 4,36,37 Inflammation can further affect and damage blood circulation into the disc, thereby impairing oxygen and nutrient supply. As shown in the degeneration progression diagram (Figure 9), inflammation is responsible for irreversible damage of parts of the vertebral body and endplate structure, which is essential for nutrient vehiculation to the disc. As a consequence, the oxygen supply and nutrient availability to the disc are inadequate, and disc degenerates.
Hence, in this study, inflammation was used to induce IVDD, which ensured the success rate of degeneration and shortened the modelling period. Importantly, in order to maintain intact the IVD structure, the trigger point of inflammation was moved to the vertebral body (Figure 1), and LPS used to trigger inflammation. Many studies demonstrated that LPS is a stable, long-lasting and efficient Radiological images of VI-IVDD rats show a degenerated IVD segment characterized by a reduced intervertebral space height and low MRI T2 signal. 8 According to our previous studies, the molybdenum-target plain X-ray is more powerful than common plain X-ray, in terms of image contrast and resolution of microstructures, such as caudal intervertebral space in rats. 22 Molybdenum-target plain X-ray also permitted a better localiza- The IVD collagen maturity was analysed by Masson staining ( Figure 4). 35 We also found that, in the VI-IVDD model, several immature fresh collagen fibres formed in the IVD, which degenerated over time. Collagen fibre formation gradually occurred from week 1 to 3, to reach a maximum at about week 3. Afterwards, the newly formed collagen fibres matured at 4 weeks.
As a whole, these data show that the process of collagen tissue remodelling is associated with IVDD, a condition that, in VI-IVDD model, occurs when IVDD reaches a peak plateau stage (P-stage) at about week 4.
Aggrecan and Col-II expressions are high in NP and AF of normal IVDs. 8 IHC ( Figure 5) showed that, in the VI-IVDD model, Aggrecan and Col-II decreased in NP and AF, but increased in EP. This condition could be related to the remodelling and degeneration of the damaged tissue. 41 In addition, in our VI-IVDD model, MMP-3 expression remarkably increased, together with fresh collagen fibres, as assessed by Masson staining, suggesting that the extracellular matrix was subjected to remodelling. 42 Hence, it seems that the IVD in the VI-IVDD model recapitulates the natural degeneration process. In addition, we found that the inflammation spreading is strictly time-dependent ( Figure 9). In fact, during the early phase of inflammation, inflammation signs were limited to homolateral vertebral bodies and reached the contralateral vertebral bodies in the late phase of inflammation. In the late phase of inflammation, IVDD achieved the P-stage discussed above. In the NIVD group, early-tomiddle transition phase, as well as P-stage achievement, was shorter than in the Mid group since the trigger point of inflammation was closer to homolateral IVD. Compared with our model, in the previously mentioned IVD damage model, the degeneration process may initiate starting directly from the middle phase, so that the associated degeneration process is too fast to be studied appropriately ( Figure 9B).
The cGAS/STING signalling pathway has been closely related to the activation of inflammatory responses. Cytoplasmic free DNA is recognized as a danger signal by DNA receptor cGAS.
The downstream STING acts as an adaptor molecule and then activates the downstream signals TBK1 and IRF3 to produce cytokines stimulating specific immune responses. 44 Previous studies demonstrated that LPS can induce the production of several inflammatory cytokines by activating the cGAS/STING signalling pathway 25 and that inflammatory responses activated by the cGAS/STING signalling pathway can in turn induce IVDD. 45 In this study, we found that cGAS/STING signalling pathway activated by LPS is closely related to IVDD. Figure 10 shows the activation of the cGAS/STING signalling pathway in the VI-IVDD model and identifies this molecular cascade as the potential mechanism inducing IVDD ( Figure 10D). In our model, we hypothesize that LPS is responsible for massive M1 macrophage infiltration and for tissue damage and inflammation. The latter lead, in turn, to cell death and release of free DNA, also from mitochondria, that could trigger downstream signals, such as STING, TBK1 and IRF3, by

| CON CLUS IONS
In this study, we describe a new type of vertebral inflammationinduced caudal IVDD rat model (VI-IVDD), as well as the processes of inflammation and progression of IVDD condition (Figure 9). Our LPS-inducible model is strictly associated with cGAS/STING signalling pathway activation ( Figure 10). Compared with other commonly used IVDD models, this model does not determine direct damage to the IVD structure, which is convenient for studying the mechanism of natural IVDD. Moreover, it has the advantage of being easily handling, having low costs and high success rate, and short modelling period, and achieving controllable degenerative segment and degeneration speed (Mid group: basic IVDD, NIVD group: accelerated IVDD). The effectiveness and repeatability of the VI-IVDD model were verified by imaging, histomorphology, histochemistry, cytochemistry (inflammation and apoptosis) and molecular mechanism.
Our model could have broad application prospects and could be helpful for unveiling new pathological mechanisms underlying IVDD condition.

ACK N OWLED G EM ENTS
We acknowledge the financial support from the Multicenter Clinical

CO N FLI C T O F I NTE R E S T S
The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are included in this published article and available from the corresponding author upon reasonable request.