Introduction

Low back pain (LBP) is a leading cause of global disability [1] and current treatments either lack efficacy or are limited by undesired side-effects. Degeneration in intervertebral discs (IVDs) is one of main causes of LBP and is often diagnosed by MRI assessment. Patients often incorporate lifestyle-based interventions into their treatment such as running, yoga and aerobic fitness, all of which are shown to attenuate pain, disability or both [2]. While beneficial effects of physical activity in LBP are well documented, its impact on IVDs is not well explored.

In the present study, we used our previously reported mouse model of chronic LBP induced by a single puncture of the mouse L4/5-IVD [3, 4] to investigate the evolution of lumbar IVD degeneration over time using histological and radiologic tools and the relationship between degeneration and behavioural changes. The effect of long-term physical activity on behaviour, degeneration and innervation of injured and adjacent, non-injured IVDs was also investigated.

Materials and methods

Please see supplementary methods for additional information on animals, surgical, behavioural and radiologic procedures.

Animals

Four cohorts (i.e. 0.5 M, 3 M, 12 M, 12 M + run) of female 20 g CD1 mice were included in this study (Charles River, Canada, 3/cage), see Table 1. Behavioural and radiologic assessments and tissue harvest were performed 0.5-, 3- or 12-months following L4/5-IVD injury. 12 M + run cohort: starting 2 weeks post-injury, active mice received a home cage running wheel (dome + wheel, Bio-serve, Flemington, NJ, USA). 12 M-cohort: sedentary mice received a sham device (the dome and wheel were fixed together to prevent rotation and voluntary activity).

Table 1 Number of animals per cohort and assay
Full size table

Mice injured at levels other than L4/5 were excluded from the study (Table 1). All experiments were approved by the Institutional Animal Care and Use Committee of McGill University and conformed to national and international ethical guidelines [5].

Surgical procedure

Injury to the L4/5-IVD was induced as described previously [3] by a single ventral mechanical puncture with a 30G needle equipped with a stopper (0.7 mm). Allocation to the sham or injured surgery was randomly assigned across cage mates using a Latin square design.

Behavioural procedures

General health

Locomotor capacity (rotarod) Latency to fall in seconds was measured using an accelerating rotarod (IITC) with the mouse adapter (rod diameter, 3.2 cm) [6].

Voluntary running Assessed using wireless Counting Wheels (ENV-047, Med associates Inc.) for 60 min.

Axial discomfort

Grip force assay Used to assess tensile strength (tolerance to axial stretching), the forearm grip strength was measured using a Grip Strength Metre (Stoelting Co, Wood Dale, IL, USA) [6].

Tail suspension assay Mice were individually suspended underneath a platform by the tail, videotaped for 180 s. Duration in immobility was quantified [6].

FlexMaze assay Mice freely navigated a corridor with staggered doors (8 × 80 cm, 15 min) that required lateral flexion [6].

Radiating pain

Both hindpaws were assessed with a 15-min interval and averaged.

Cold sensitivity The total duration of acetone-evoked behaviours (e.g. flinching, licking or biting) was measured after a drop (25 μL) that was applied to the plantar surface [6].

Mechanical sensitivity von Frey filaments (Stoelting Co., Wood Dale, IL, USA) were applied to the plantar surface using the up-and-down method as previously described [6].

Disc degeneration assessment

Lumbar spine X-ray

Disc height index (DHI) was measured using lateral X-ray images of the spines (Faxitron MX-20 X-Ray LLC, Lincolnshire, IL, USA) performed under deep anaesthesia [4].

Lumbar spine magnetic resonance imaging: preparation, scans, and quantification

After vascular rinse and fixation (4% PFA, intra-cardiac), the T1 to S4 spinal segment was harvested, post-fixed (48 h), embedded (15 mL tube, 10% agarose) and scanned using Bruker standard sagittal RARE-T2 images on a Biospec 70/30 (7 T, Avance III) with Paravision 5.1 (Bruker BioSpin Corporation, Billerica, MA, USA) [3]. Scans were analysed using MIPAV software package, version 7.1.1. using the median section for each lumbar disc. Each region of interest (ROI) included cartilaginous endplates, annulus fibrosus (AF), nucleus pulposus (NP) and extruded material for bulging discs.

Histological assessment

After completion of MRI, samples were decalcified (4% EDTA), cryoprotected, embedded in OCT (Tissue-Tek) and sectioned in the sagittal plane (16 μm) using a cryostat.

Histological staining

FAST staining was performed as previously described [4].

Immunofluorescence histochemistry

Immunohistochemistry was performed using rabbit anti-PGP9.5 (1:1000, Ultraclone, RA95101) and sheep anti-CGRP (1:2000, BML‐CA1137‐0100; Enzo Life Sciences) as previously described [4]. Images were quantified using ImagePro on three non-consecutive sections and averaged for each disc of each animal.

Data analysis

Comparisons between sham and injured mice of each cohort were performed using 2-way ANOVA followed by Fisher’s LSD using Graphpad Prism 8.0 software. Data are expressed as mean ± SEM. All experiments were performed blind to the surgical status.

Results

Time-course of behavioural changes following L4/5-IVD puncture injury

Mice with L4/5-IVD injury had similar (0.5- and 3-months) or increased weight (12 months) compared to their sham controls. The increase at 12 months post-injury was reversed by long-term physical activity (Fig. 1a). While no changes in locomotor capacity were associated with L4/5-IVD injury in the rotarod assay, active animals (sham and injured) had increased capacity compared to age-matched sedentary animals (Fig. 1b). L4/5-IVD injury resulted in a delayed decrease in voluntary wheel running at 12-months that was not observed at earlier time points; this deficit was blocked by physical activity (Fig. 1c). Axial discomfort was assessed using the grip force, tail suspension and FlexMaze assays (Fig. 1d–f and Supplementary Fig. 1). These behavioural assays discriminated between animals with spontaneous or induced IVD degeneration [3, 6,7,8,9], but no persistent differences were observed in the current study. Increased activity had no major impact on axial discomfort, with the exception of the increased grip strength of active injured mice 12 months post-injury (Fig. 1d).

Fig. 1
figure 1

Long-term behavioural consequences of L4/5-IVD injury in active and sedentary mice. 0.5-, 3- or 12-months following L4/5-IVD injury (dark grey) or sham surgery (light grey), cohorts of sedentary and active mice were evaluated for general health (a body weight (in grams), b locomotor capacity (in seconds) in the rotarod assay, c voluntary running capacity assessed as the number of revolutions performed on a wireless Counting Wheels over 1 h), axial discomfort (d Resistive force (in grams) during the grip force assay, e percentage of time in immobility during the 3-min tail suspension assay, f tolerance to lateral bending measured as change in exploration speed in the FlexMaze assay), and radiating discomfort (g cold sensitivity assessed as time spent in acetone-evoked behaviour (sec), h mechanical sensitivity measured as 50% withdrawal threshold (in grams) in the von Very assay). 2-way ANOVA follow by Fisher's LSD; *p < 0.05, **p < 0.01. N = 5–14

Full size image

Development and maintenance of signs of radiating pain were evaluated. After L4/5-IVD injury, mice demonstrated hyper-sensitivity to cooling stimuli at 3- and 12-month post-injury that was reversed in active mice (Fig. 1g). Consistent with previous studies [3], no changes in the 50% withdrawal threshold in response to mechanical stimuli were observed (Fig. 1h).

Time-course in radiographic (X-ray) and structural (T2-MRI) changes following L4/5-IVD puncture injury

X-ray

Disc injury resulted in loss of L4/5-IVD height as early as 2 weeks after puncture that persisted throughout the duration of the study. Long-term running activity protected against the loss of disc height 12 months post-injury (Fig. 2b).

Fig. 2
figure 2

Long-term histologic and radiologic consequences of L4/5-IVD injury in active and sedentary mice. a Example of histological (FAST staining), X-ray, and T2-MRI performed in the same sample (also see Fig. 3a, mouse H3). The lumbar spine used here was harvested in a sedentary mouse 12 months after. The injured L4/5-IVD shows signs of degeneration upon histological (collapsed disc with no distinctive nucleus pulposus) and T2-MRI (dark disc) examinations. In this example, the uninjured L2/3-IVD also shows a typical degenerative pattern (i.e. no distinctive nucleus pulposus upon histological analysis, and dark disc upon T2-MRI analysis). # secondary degenerative L2/3-IVD. b–d 0.5-, 3- or 12-months following L4/5-IVD injury (dark grey) or sham surgery (light grey), cohorts of sedentary and active mice were evaluated for anatomic change. b Disc Height Index was measured on X-ray images. Disc hydration (c) and compartmentalization (d) were evaluated on T2-MRI images as median signal intensity and standard deviation for each lumbar IVD. 2-way ANOVA follow by Fisher’s LSD; *p < 0.05, ***p < 0.001. N = 5–14

Full size image

T2-MRI

The signal intensity median value and standard deviation (SD/signal) were measured on T2-MRI images, and reflect disc hydration and compartmentalization both signs of disc health. In L4/5-IVD, hydration and compartmentalization were significantly decreased 2 weeks post-injury (Fig. 2c, d). While compartmentalization did not evolve further hydration was progressively reduced with age in both injured and non-injured discs. Long-term physical activity had no impact on T2-MRI measures in this study.

Impact of L4/5-IVD puncture injury on adjacent discs

To examine the impact of disc injury on adjacent lumbar discs, the SD of the signal intensity (SD/signal) for each lumbar IVD was determined. The SD/signal ranged from 1111 to 6649. For each cohort, the average (XSD/signal)-1 SDSD/signal of the L4/5-IVD measurements in the sham animals was set as the minimum value for a healthy disc. A SD/signal above this value indicates compartmentalization is intact (healthy), a SD/signal below suggests compartmentalization is compromised (unhealthy) (Fig. 3).

Fig. 3
figure 3

Injured versus secondary degenerating discs in active and sedentary mice. a Heat map of individual lumbar degeneration upon T2-MRI examination, 0.5-, 3- or 12-months following L4/5-IVD injury (dark grey) or sham surgery (light grey) in cohorts of sedentary and active mice. The scale for grey intensity was automatically adjusted for each cohort to avoid age-related bias as follows: X = (averagesham(L4/5-IVD) − 1 SDsham(L4/5-IVD)) for each cohort [X = 3062 (0.5 M-cohort), X = 3720 (3 M-cohort) and X = 4392 (12 M-cohorts)]. b, c total (PGP9.5-ir, Fig. 3b-left panel) and nociceptive (CGRP-ir, b right panel) innervation, and Disc Height Index (c) were measured in sedentary (a− c) or active (d) mice 12-months post-surgery. Measures are presented for L4/5-IVD [sham (light grey) vs. injured (dark grey) mice] or for other level (L1/2, L2/3, L3/4, L5/6) uninjured IVDs from L4/5 injured animals. The other level IVds were futher classified as Normal (textured light grey) versus Degenerative (textured dark grey) based on T2-MRI analysis. For additional details, see Supplementary Method. 2-tailed t-test; *p < 0.05, ***p < 0.001

Full size image

Using this as a measure of degeneration, we observed an increase in adjacent lumbar disc degeneration with time (Fig. 3a). This was particularly noticeable in the sedentary 12-month post-injury cohort, which presented with numerous degenerating discs in the upper lumbar levels. In the active 12-month injured cohort, the lower lumbar discs were mostly affected (Fig. 3a).

Adjacent discs: innervation and height

In the 12-month post-injury sedentary and active cohorts, lumbar IVDs other than L4/5 were categorised as normal, degenerating or intermediate using the SD of the MRI signal intensity (SD/signal, see suppl. Figure 3). Mechanically injured and uninjured IVDs with secondary degeneration had similar MRI measurements, and no effect of running was observed. In contrast, increased innervation was observed in L4/5 injured IVD but not in the secondary degenerating IVDs (Fig. 3b). Finally, when segregated based on the MRI categories, disc height in non-injured degenerating IVDs was protected by running activity (Fig. 3c).

Discussion

In mice, a single-level L4/5-IVD injury produced robust signs of radiating pain that were alleviated in animals with unlimited access to a home cage activity wheel. By 12-months post-injury, disc degeneration spreads to other levels throughout the lumbar spine. Those secondary degenerating discs presented radiologic signs of degeneration but lacked the pathological innervation observed in the directly injured IVDs. We propose the secondary discs should be further investigated as they may represent a clinically-relevant aspect of human disc degeneration.

Single L4/5-IVD injury and signs of discomfort in mice

Mechanical L4/5-IVD injury did not result in persistent changes in general health or axial discomfort for up to one year following injury. However, consistent with previous studies [3], signs of radiating pain (hyper-sensitivity to hindpaw cooling) were observed and were reversed by long-term physical activity.

Axial discomfort suggestive of discogenic pain

The lack of behavioural signs of axial pain in this study is in conflict with previous report on L4/5-IVD injury-induced pain behaviour [3], where behavioural signs of axial discomfort peaked 3 to 9 months after injury. While it is possible the current study missed the critical time window, it is also clear that assessment of axial discomfort in rodents is a major challenge [10]. This is further amplified by the lack of consensus on disc injury models and the behavioural methods used (i.e. altered physical function [11, 12], rearing [6, 13], grip test or tail suspension [14]) that transfer poorly from one model to another, even within a single team. Since the utility of clinically-relevant animal models of back pain depends on our ability to assess axial discomfort, new creative approaches to behavioural assays are desperately needed [10].

In contrast to the behavioural measures, radiologic (X-ray), structural (MRI) and physiological (innervation) consequences of disc injury are stable and reproducible. This mismatch highlights the poor correlation between the severity of the disease and the behavioural changes observed in animals. Thus, in rodents, as in humans, back pain is not always synonymous with disc degeneration [15].

Hindpaw hyper-sensitivity suggestive of radicular pain

While axial discomfort remains technically challenging to evaluate in animals, most studies investigating behavioural consequences of disc injury/degeneration report hyper-sensitivity to stimuli applied to hindpaws. In our initial publication [3], we observed that cold hyper-sensitivity developed in a biphasic fashion with an early, transient peak at 0.5-months post-injury that resolved by 6-months and re-emerged at 12-months. Here we observed a delay in the onset of cold hyper-sensitivity at 3- and 12- but not 0.5-months post-injury.

In contrast to the delayed onset of radicular pain, the severity of degeneration of the injured discs by both X-ray and MRI was fully established by 0.5-months and persisted throughout the study, suggesting a mismatch between radiating pain and degeneration severity. In primary care, approximately 60% of patients with low back pain also report pain in one or both legs [16]. Interpedicular height has been suggested as a predictor of radicular pain in adult degenerative scoliosis [17] and mechanical radiating pain co-varies with severity in disc narrowing in a rat model of disc injury [18]. In our previous studies with both induced and progressive models of mouse IVD degeneration, disc height has been consistently associated with cold hyper-sensitivity on the hindpaw [3, 6, 7].

In the current study, long-term access to voluntary running wheels was used to model an active lifestyle and prevented signs of radiating pain 12-months post-injury. Activity also attenuated the loss of disc height in secondary degenerating IVDs, furthering the argument for a direct relationship between pathologic disc narrowing and radiating discomfort.

Progression of disc degeneration

We performed radiologic and histological analysis of the 5 lumbar discs for each animal, including both injured and uninjured L4/5 and adjacent segments of the spine.

Progression of disc degeneration over time

We previously reported increased innervation and severe degeneration by histological assessment in injured IVDs as early as 0.5 M post-injury that persisted for up to one year [4]. In the current study, we demonstrate that degeneration severity of injured L4/5-IVDs was similar from 0.5 to 12 months by radiologic (X-ray and MRI) measurements. These data suggest that structural damage to the injured IVD does not evolve further after the acute injury. Long-term physical activity had no impact on radiologic measures of the injured IVD. However, since non-injured IVDs narrow with age, the differences between injured and non-injured IVDs faded with time.

Progression of degeneration in adjacent segments over time

A major finding of this study is that disc degeneration spreads to adjacent discs with time in mice. The secondary degenerating discs had reduced height and altered MRI signal of similar magnitude to the primary injured disc; radiologically, it would be difficult to distinguish between mechanically injured and non-injured secondary degenerating IVDs 12-months post-injury. In contrast, these secondary degenerating discs were not abnormally innervated. Long-term physical activity prevented the secondary degenerating discs from narrowing but had no impact on MRI signal intensity.

The adjacent segment spread is clinically-relevant and is well documented in humans both spontaneously and after spinal fusion [19,20,21] or vertebral fracture [22]. It has also been described after spinal fusion in animals such as goats [23] or ovariectomized rats [24]. To our knowledge, this is the first time that the degeneration of discs other—and sometimes even distant—than the mechanically damaged one has been documented, without spinal fusion, in mice. This clearly calls into question the validity of using adjacent discs as controls in molecular, physiological or biomechanical studies in animals.

In this model, the spreading of degeneration cannot be attributed to genetic predisposition or congenital weakness. Rather, it is a downstream consequence of the initial L4/5 injury, perhaps related to spinal instability. Additional studies of these secondary degenerating discs are needed, including assessment of sympathetic innervation, vascularization, pro-inflammatory and pro-nociceptive factors (e.g. TNFa, IL8, NGF) and extracellular matrix remodelling. We suggest that secondary degenerating discs represent a clinically-relevant aspect of degeneration of human discs, especially in cases where multiple levels are affected.

Conclusions

Here we report the spread of degeneration to other lumbar discs following L4/5-IVD injury. The secondary degenerating discs lacked the pathological innervation observed in the directly injured IVDs. Increased physical activity attenuated behavioural signs of radiating pain and prevented the loss of disc height in non-injured, secondary degenerating discs. These studies support physical activity-based interventions for discogenic low back pain and call for additional focus on adjacent segment degeneration, which may represent a clinically-relevant model of the human condition.