Intradiscal Inflammatory Stimulation Induces Spinal Pain Behavior and Intervertebral Disc Degeneration In Vivo

Abstract

Degeneration of the intervertebral disc (IVD) results in a range of symptomatic (i.e. painful) and asymptomatic experiences. Components of the degenerative environment, including structural disruption and inflammatory cytokine production, often correlate with pain severity. However, the role of inflammation in the activation of pain and degenerative changes has been complex to delineate. The most common IVD injury model is puncture; however, it initiates structural damage that is not representative of the natural degenerative cascade. In this study, we utilized in vivo injection of lipopolysaccharide (LPS), a pro-inflammatory stimulus, into rat caudal IVDs using 33G needles to induce inflammatory activation without the physical tissue disruption caused by puncture using larger needles. LPS injection increased gene expression of pro-inflammatory cytokines (Tnfa, Il1b) and macrophage markers (Inos, Arg1), supported by immunostaining of macrophages (CD68, CCR7, Arg1) and systemic changes in blood cytokine and chemokine levels. Disruption of the IVD structural integrity after LPS injection was also evident through changes in histological grading, disc height, and ECM biochemistry. Ultimately, intradiscal inflammatory stimulation led to local mechanical hyperalgesia, demonstrating that pain can be initiated by inflammatory stimulation of the IVD. Gene expression of nociceptive markers (Ngf, Bdnf, Cgrp) and immunostaining for neuron ingrowth (PGP9.5) and sensitization (CGRP) in the IVD were also shown, suggesting a mechanism for the pain exhibited. To our knowledge, this rat IVD injury model is the first to demonstrate local pain behavior resulting from inflammatory stimulation of caudal IVDs. Future studies will examine the mechanistic contributions of inflammation in mediating pain.

Graphical Abstract

A new rat model of intervertebral disc (IVD) degeneration initiated by inflammatory stimulation using intradiscal injection of lipopolysaccharide (LPS) was developed. It decreased structural integrity of the IVD, increased local and systemic inflammatory cytokines, and increased macrophage presence in the disc. In vivo intradiscal inflammatory stimulation also increased expression of neurons, nociceptive markers, and localized spinal pain behavior.

Introduction:

Low back pain is the leading cause of disability and is often associated with degeneration of the intervertebral disc (IVD)^1,2. IVD degeneration is associated with pain ranging in severity from asymptomatic to severe and manifests as extracellular matrix (ECM) disruption³. ECM structural changes have been previously associated with pro-inflammatory responses characterized by sustained, elevated levels of many cytokines, including tumor necrosis factor alpha (TNFα) and interleukin 1 beta (IL-1β) in patient IVD samples⁴. Furthermore, cytokine profiles have been correlated with degree of disc degeneration, showing significantly different expression levels between degenerated intact and herniated IVDs^5–7. Interestingly, gene expression of Tnfa and Il1b has been correlated with nerve growth factor (Ngf), a nociceptive marker known to be increased in symptomatic (i.e. painful), degenerating IVDs^8,9. Degeneration grade and protein expression of neuron markers, calcitonin gene-related peptide (CGRP) and protein gene product 9.5 (PGP9.5), have also been correlated¹⁰. These findings suggest a causative relationship may exist between inflammatory cytokines and discogenic pain.

There is extensive evidence for a role of pro-inflammatory cytokines in certain disc diseases. For example, herniated discs exhibit immune cell infiltration and increased pro-inflammatory cytokines, as well as radiating pain symptoms caused by mechanical compression and chemical irritation from the herniated tissue^11–14. However, the role of inflammation in triggering localized pain of discogenic origin due to a contained but degenerated disc is less understood.

Needle puncture or disc lesion injuries are the most popular models used to simulate disc degeneration in vivo, and have previously been implemented in both the lumbar and caudal spinal regions of mouse, rat, and rabbit models. These studies demonstrated that puncture or lesion successfully reproduces morphological changes to the IVD and behavior attributed to the experience of pain in animals. After lumber disc injury, both mechanical and thermal sensitivity were observed with peaks occurring at 3-9 months and 9-12 months post-injury, respectively^14–19. However, inflammatory changes in these models are thought to be caused primarily by physical tissue disruption leading to an acute and transient inflammatory response post-injury. This differs from the presence of sustained, chronic inflammation and pain seen in human disc degeneration^20–22.

To assess the potential of inflammatory stimulation in contributing to pain behavior, previous studies evaluated intradiscal injection of TNFα or other growth factors, including vascular endothelial growth factor (VEGF) and nerve growth factor (NGF), in conjunction with needle puncture injury, and found accelerated hind paw mechanical sensitivity compared to disc puncture injury alone²³. Injection of TNFα in rat and porcine models also resulted in increased histological score and intradiscal TNFα expression, which were predictive of pain behavior²⁴. Interestingly, anti-TNFα treatment was shown to prevent changes in degeneration grade and mechanical sensitivity compared to sham levels²⁵. While these studies demonstrate that intradiscal inflammation can accelerate radiating pain in the paw caused by needle puncture, it remains unknown whether activation of inflammatory signaling independent of tissue trauma can trigger localized spinal pain of discogenic origin.

Recent studies have highlighted the role of innate immune activation, particularly that of toll-like receptors (TLRs), in the pathogenesis of disc degeneration^26–28. TLRs are members of a receptor family activated by damage associated molecular patterns (DAMPs), such as fibronectin fragments or high mobility group box 1 (HMGB1) protein, resulting in persistent pro-inflammatory signaling and degenerative effects^29–31. LPS has been shown to activate TLR4 in vitro inducing a pro-inflammatory cascade in IVD cells^32,33. Further, injection of LPS directly into the nucleus pulposus (NP) space of rat caudal IVDs has caused moderate degenerative changes in the IVD, with increases in tissue levels of IL-1β, TNFα, HMGB1, and macrophage migration inhibitory factor (MIF)²⁸. Animal models of disc degeneration, have also shown an association between increased TLR4 expression and pain, while inhibition of TLR4 decreased pain and pro-inflammatory cytokine production^34–36. However, the extent to which inflammatory signaling resulting from TLR activation causes pain behavior requires further investigation.

The goal of this study was to evaluate behavioral pain responses of rats after inflammatory activation of caudal IVDs by LPS injection in vivo. A range of LPS dosages were initially investigated to identify a dose causative of matrix integrity disruption. The behavioral responses of rats to mechanical and thermal stimuli were evaluated locally in the spine and distally in the hind paw. Caudal IVDs, as opposed to lumbar, were selected to create an easily reproducible model with a less invasive surgical procedure that does not require disruption of adjacent tissues to access the IVD space. Most prior studies have focused on the lumbar spinal region with IVD puncture producing innervation and mechanical hyperalgesia^10,14,18. To our knowledge, the current study is only the second to demonstrate pain behavior after injury to the caudal IVD, and the first to evaluate pain in an inflammatory injury model¹⁷. Demonstration of structural changes, a sustained inflammatory profile, and a local mechanical pain phenotype resulting from only inflammatory stimulation fills a critical gap in the area of animal models of disc degeneration.

Materials and Methods:

Surgical procedures were performed separately on 4 cohorts of animals (N=54 in total), as described below. Analgesia was not administered to prevent potential interference with pain outcomes. Institutional animal care and use committee (IACUC) approval was obtained prior to the start of experiments.

Surgical Procedure – Cohort 1: Dose Response

Male Sprague Dawley rats (N=18 total) weighing 300-350 g, corresponding to approximately 3-month-old rats that are skeletally mature, were anesthetized with 3-5% isoflurane and 3-4 L/min oxygen. Once anesthetized, isoflurane was lowered to 1-3% and oxygen to 1 L/min, before an incision was made exposing 4 caudal or coccygeal (Co) motion segments, approximating levels Co3-4 to Co6-7. Phosphate buffered saline (PBS) or an LPS (Sigma, Cat#. L2630) solution was injected into the NP space of all exposed discs as previously described by Rajan et al²⁸. Briefly, a 33 gauge (G) needle was inserted 4mm into the center of the disc with clamp guidance. The needle size was chosen as it is <25% of the average caudal disc height for all disc levels used in this study, minimizing tissue damage^37,38. LPS was sonicated for 30 minutes before being diluted to the desired concentration in saline. A single dose (2.5 μl) of saline (n=18), as a sham control, or LPS (either 1, 10, or 100 μg/mL, n=6 for each dose) was injected slowly over 60 seconds into alternating discs using a microliter syringe (Hamilton, Cat#. 65460-06). On an additional animal, only an incision was made without needle insertion functioning as an incision only control. The incisions were closed with 4-0 nylon sutures (DemeTECH) and animals were allowed unrestricted activity until the end of the study at 2-, 7-, or 28-days post-injury. After dissection, samples from this cohort were analyzed for disc dimensions, biomechanical properties, and biochemical composition.

Surgical Procedure – Cohorts 2 & 3

Male Sprague Dawley rats (Cohort 2: N=9 total, Cohort 3: N=15 total) were anesthetized with isoflurane, and an incision was made in the caudal spine exposing 3 motion segments, approximating levels Co4-5 to Co6-7. A 33 G needle was inserted 4 mm into the center of each disc with clamp guidance. 2.5 μl of USP 0.9% saline (Cohort 2: n=4, Cohort 3: n=5) or LPS (100 μg/mL, Cohort 2: n=5, Cohort 3: n=5) was injected slowly over 60 seconds into the exposed caudal discs of each animal using a microliter syringe, with all discs in an animal receiving the same treatment. On a set of rats in Cohort 3, only an incision was made without needle insertion as an incision only control (n=5). Incisions were closed with 4-0 nylon sutures and animals were evaluated longitudinally for pain behaviors and blood serum cytokine levels until 28 days after injury.

Surgical Procedure – Cohort 4

Male Sprague Dawley rats (N=12 total) were anesthetized with isoflurane. Using fluoroscopic guidance, caudal motion segments Co5-6 to Co8-9 were identified and exposed. 4-0 nylon sutures were used to mark above and below the IVDs of interest. 2.5 μl of saline (n=6) or LPS (100 μg/mL, n=6) was injected into all exposed discs of each animal, the incision was closed as described above, and animals were euthanized at 14 or 28 days post-injury. Samples from cohort 4 were collected for disc height analysis, histological grading, gene expression, and immunostaining.

Dimensions and Biomechanical Testing –

Bone-disc-bone motion segments were collected from cohort 1 and the surrounding connective tissue was removed. Motion segment height and diameter was measured with digital calipers. Using the measured values, cross sectional area (CSA) and aspect ratio were calculated using the following formulas:

$$CSA=\pi {\left(\frac{d}{2}\right)}^{2}AspectRatio=\frac{h}{d}$$

where d represents diameter and h represents sample height. Data was normalized to the corresponding sham from the same animal. Differences in sample geometry were compared in LPS versus sham groups using Student’s t-tests.

For mechanical testing, motion segments were subjected to unconfined compression between 2 stainless steel platens while submerged in PBS. Using the Instron testing frame (Instron 5566) equipped with a 10 N load cell, samples were first pre-loaded with a 0.1 N tare load followed by cyclic loading to 3 N applied at 0.1 Hz for 30 cycles. Disc segments were then subjected to a creep load of 3 N until equilibration. The resulting displacement was measured and creep strain was determined. The dynamic and equilibrium moduli were then calculated using the following formulas:

$$DynamicModulus=\frac{\frac{Force}{\Delta Displacement}*h}{CSA}EquilibriumModulus=\frac{\frac{Load}{CSA}}{CreepStrain}$$

After mechanical testing, disc segments were stored at −20°C for analysis of biochemical content. Differences in biomechanical properties of the LPS and sham groups or between the 14- and 28-day time points were determined using Student’s t-tests. Differences between sham and the incision only control 28 days after injury were also calculated using Student’s t-tests. Data is presented as mean ± standard deviation.

Biochemical Content –

Individual discs were thawed, bone segments were removed, and annulus fibrosus (AF) and NP were separated using a biopsy punch. Tissue wet weights were measured, samples were dried in a vacuum desiccator, and lyophilized tissues were weighed again to obtain tissue dry weight. Water content was calculated based on percent differences in wet and dry weights. Samples were then digested overnight at 60°C in 20 μL papain aqueous suspension (≥16 units/mg, Sigma-Aldrich) per 1 mL solution in papain buffer. Tissue digests were analyzed for DNA content using a Picogreen assay (Invitrogen), glycosaminoglycan (GAG) content using the Blyscan assay (BioVendor), and collagen content using the hydroxyproline (OHP) assay³⁹. Biochemical content was reported as concentrations and normalized to tissue wet weight. Outliers were detected and removed using the ROUT method before differences in the biochemical content of sham and LPS IVDs were evaluated using one-way analyses of variance (ANOVAs) with Fisher’s least significant difference (LSD) post-hoc tests, to compare each LPS dose versus sham⁴⁰. Differences between sham and control were determined using Student’s t-tests. Data is presented as mean ± standard deviation.

Behavioral Testing –

Behavioral testing was performed on rats from cohorts 2 and 3. Animals in cohort 2 studies underwent the Pressure Algometry Measurement (PAM) test in the dorsal tail, von Frey test in the hind paw, and tail flick test in the distal tail. Animals in cohort 3 studies underwent the PAM test in the dorsal tail, von Frey test in the ventral tail, Hargreaves test in the tail base, and tail flick test in the distal tail. Results from cohort 2 and 3 were analyzed together for behavior tests repeated independently in both cohorts. Behaviors were evaluated at day 0, as a baseline, and then at 1, 7, 14, 21, and 28 days post-injury. Animals were acclimated to each test’s equipment and the experimenter in a quiet room before testing was conducted. All behavioral testing was blinded.

Mechanical hyperalgesia was measured using the von Frey and PAM tests. The von Frey test was performed on each animal in one of 2 locations: hind paw or ventral tail^17,41. Rats were placed in individual cubicles on top of a suspended wire mesh surface and allowed to acclimate to the testing space for 15-20 minutes. After acclimation, von Frey filaments (Ugo Basile, Cat#. 37450) were pressed perpendicularly to the plantar surface of the selected hind paw (cohort 2 studies) or ventral tail (cohort 3 studies), where they were held in position for approximately 2-3 seconds with enough force to cause a slight bend in the filament. Testing was started with a 2 g force filament and the response of the rat was evaluated. Positive responses included an obvious withdrawal of the hind paw or tail from the filament and/or flinching behaviors and licking. If a positive response was observed, a lower force filament would be tested, while no response would be followed by testing with a higher force filament. Data was collected until 3 positive responses were observed or the maximum force threshold (15 g) was reached without achieving a response. The average force from positive responses or the maximum force threshold was recorded and normalized to the baseline average in each group.

The PAM procedure was adapted from Kim et al. and was conducted using an Ugo Basile instrument (Cat#. 38500) accompanied by a force transducer (Cat#. 58500-2)⁴². The rat was manually restrained by the experimenter using a padded glove and/or towel, and allowed to acclimate to the holder until resistance ceased. The PAM transducer was placed on the testers thumb and slowly pressed on the dorsal skin aligned with the experimental discs. The force was slowly increased at 100 g/second until a vocalization or physical response was observed. Following response, the test was repeated and the 2 associated forces were averaged and recorded. A cutoff of 1500 g applied force was set to avoid tissue damage. Recorded forces were normalized to the baseline average force within each group for the cohorts independently. After normalization, data from cohorts 2 and 3 were analyzed together.

Thermal hyperalgesia was measured using the tail flick and Hargreaves tests according to the methods described by Mohd Isa et al¹⁷. The tail flick test was conducted using the Ugo Basile instrument (Cat#. 37360). Rats were previously acclimated to the handler, therefore no additional acclimation was necessary. The rat was manually restrained by the experimenter and held steady on the top of the instrument. Radiant heat was focused 4-7 cm from the distal end of the tail. Heat was applied until the rat responded by flicking the tail away from the heat source, and the response time was recorded. The infrared (IR) intensity was set at 20 with a cutoff time of 15 seconds to prevent the potential for tissue damage.

For the Hargreaves test, rats were placed in the Hargreaves arena (Ugo Basile) and allowed to acclimate for 15-20 minutes. Heat was applied ventrally at the base of the tail with withdrawing, flinching, licking, or biting considered a positive response. The time until display of a positive response was recorded as the withdrawal time. IR intensity was set to 40 with a cutoff time of 20 seconds. In all behavioral tests a two-way ANOVA was used with group (Sham vs. LPS or Control vs. Sham) and time as variables. Between group differences at each time point and change relative to baseline within each group were evaluated using a Fisher’s LSD post-hoc test. Data is presented as mean ± standard deviation.

Disc Height Measurement and Histological Analysis –

Fluoroscopic disc height measurements and histological analysis were performed on disc segments from rats in cohort 4 studies. Fluoroscopic images were taken of all IVDs of interest pre-injury, and 14 or 28 days post-injury. Height measurements at 3 points along the diameter of the IVD (D1-D3) and each of the adjacent vertebral bodies (V1-3, V4-6) were used to calculate the disc height index (DHI) at each level using the following formula:

$$DHI=\frac{2\left(D1+D2+D3\right)}{V1+V2+V3+V4+V5+V6}$$

Post-injury DHIs were normalized to corresponding pre-injury DHIs to elucidate change in disc height. Statistical significance was determined using a two-way ANOVA between conditions and time points with a Šídák post-hoc test. Data is presented as mean ± standard deviation.

For histological analysis, disc segments were dissected 14 or 28 days post-procedure and submerged in 4% paraformaldehyde (PFA) for 48 hours. Following fixation with PFA, disc segments were washed 3 times with PBS and decalcified in 14% ethylenediaminetetraacetic acid (EDTA) at 4°C for approximately 2 weeks. After decalcification, they were again washed with PBS and transferred to 70% ethanol for transportation to the Molecular Pathology Shared Resource (MPSR) facility at the Columbia University Herbert Irving Comprehensive Cancer Center (HICCC) where paraffin embedding, sectioning, and Safranin O-Fast Green staining was performed. Stained sections were imaged and histological grading was conducted using a consensus, rat-specific grading scale⁴³. Eight categories were scored from 0, denoting healthy, to 2, representing most degenerated, by 2 blinded graders. Average scores for each category were determined, and the summed total histological grade (Maximum=16) was calculated for each disc. Statistical significance was determined using a two-way ANOVA between conditions and time points with a Šídák post-hoc test. Data is presented as mean ± standard deviation. The distribution of histological grades in each subcategory was plotted separately and significance of average scores between sham and LPS groups was calculated using Student’s t-tests.

Gene Expression Analysis –

Gene expression analysis was performed on IVDs from cohort 4 at 14 or 28 days after injury. Discs were isolated from the caudal spine with the AF and NP manually separated using a scalpel. Each region was individually snap frozen in cryogenic vials. Tissue was pulverized using a bead homogenizer and cells were lysed with TRIzol and chloroform prior to phase separation. A 1:1 ratio of 100% isopropanol and high salt solution composed of 1.2 M sodium chloride and 0.8 M sodium citrate in RNAse-free water was utilized for precipitation of RNA before purification with the RNeasy Mini Kit (Qiagen) according to the manufacturer’s protocol. Expression of genes was measured using reverse transcription-quantitative polymerase chain reaction (RT-qPCR) to evaluate changes in macrophage markers (Inos, Arg1), inflammatory cytokines (Tnfa, Il1b), and nociceptive markers (Cgrp, Ngf, brain-derived neurotrophic factor (Bdnf)). Primer sequences can be found in Table 1. Values were excluded if amplification did not occur at the same temperature or if indicated as an outliers by the ROUT method⁴⁰. Statistical significance was determined using a two-way ANOVA with Fisher’s LSD post-hoc test. Data is presented as mean ± standard deviation.

Table 1.

Primer sequences for macrophage markers (Inos, Arg1), inflammatory cytokines (Tnfa, Il1b), and nociceptive markers (Cgrp, Ngf, Bdnf).

Gene	Primer Sequence
Gapdh - FWD	5’ – GCA AGG ATA CTG AGA GCA AGA G – 3’
Gapdh - REV	5’ – GGA TGG AAT TGT GAG GGA GAT G – 3’
Inos - FWD	5’ – AAC CCA AGG TCT ACG TTC AAG – 3’
Inos - REV	5’ – GCA CAT CGC CAC AAA CAT AAA – 3’
Arg1 - FWD	5’ – CCA AGC CAA AGC CCA TAG A – 3’
Arg1 - REV	5’ – CCA GGC CAG CTT TCC TTA AT – 3’
Tnfa - FWD	5’ – CCC AAT CTG TGT CCT TCT AAC T – 3’
Tnfa - REV	5’ – CAG CGT CTC GTG TGT TTC T – 3’
Il1b - FWD	5’ – TCT GAC AGG CAA CCA CTT AC – 3’
Il1b - REV	5’ – CAT CCC ATA CAC ACG GAC AA – 3’
Cgrp - FWD	5’ – CAC GTA CAC ACA AGA CCT CAA – 3’
Cgrp - REV	5’ – CTC CAA GTC CTT GGC CAT ATC – 3’
Ngf - FWD	5’ – CTC CAA GCA CTG GAA CTC ATA C – 3’
Ngf - REV	5’ – CAC ACG CAG GCT GTA TCT ATC – 3’
Bdnf - FWD	5’ – TGG CTC TCA TAC CCA CTA AGA – 3’
Bdnf - REV	5’ – CGG AAA CAG AAC GAA CAG AAA C – 3’

Serum Cytokine Analysis –

Cytokine analysis was performed on blood collected from Cohort 2 rats at day 0, as a baseline, and 7, 14, 21, and 28 days after injury. Blood was collected into microcentrifuge tubes after behavior testing was completed using venipuncture of the retro-orbital vein, alternating sides at each time point. Blood was incubated at room temperature for 1 hour, centrifuged, and serum was isolated. Serum cytokine levels were measured using a Bio-Plex Pro^™ Rat Cytokine 23-Plex Assay (Bio-Rad, Cat#. 12005641), according to the manufacturer’s protocol. All longitudinal measurements were normalized to their baseline level within each animal. Between group differences at each time point and change relative to baseline within each group were evaluated using a two-way ANOVA with Fisher’s LSD post-hoc test. Data is presented as mean ± standard deviation.

Immunostaining Analysis –

Immunostaining was performed on all IVDs from cohort 4 collected and paraffin embedded for histological analysis. Paraffin residue was first removed in xylene and the tissue was rehydrated using ethanol solutions and water. Following rehydration, heat-mediated antigen retrieval was completed by incubating sections in sodium citrate solution at 95°C for 15 minutes. Blocking was done using 5% goat serum in 0.1% Triton-X 100 for 45 minutes. Primary antibodies including combinations of macrophage antibodies, CD68 (Abcam, Cat#. ab31630, 1:200), CCR7 (Abcam, Cat#. ab32527, 1:300), and Arg1 (Abcam, Cat#. ab91279, 1:250), or neuron markers, PGP9.5 (Proteintech, Cat#. 14730-1-AP, 1:500) and CGRP (Abcam, Cat#. ab81887, 1:500), diluted in the same blocking solution were applied overnight at 4°C. Blocking solution without primary antibodies was applied to a subset of sections as a control for non-specific binding. The following day, secondary antibodies, AF488 (Abcam, Cat#. ab150081, 1:200) and AF594 (Invitrogen, Cat#. A11005, 1:200) were diluted in the same blocking solution before being applied to sections for 1 hour at room temperature. A DAPI mounting media (VECTASHIELD, Cat#. H-1800) was used before coverslips were applied and slides were imaged on a Zeiss Axio Observer. Representative images were taken and CD68, CCR7, and Arg1 expression was quantified using ImageJ to obtain regional mean fluorescence intensity (MFI) for the NP, AF, endplate and growth plate (EP + GP), and tissue regions distal to the organized lamellar structure of the outer AF (periphery) Representative images showing the regions used in this analysis are shown in Supplemental Figure 3A–B. Statistical significance was determined using a two-way ANOVA with a Fisher’s LSD post-hoc for each IVD region with treatment and time as variables. Data is presented as mean ± standard deviation.

Results:

LPS Dose Response: Dimensions, Biomechanics, and Biochemical Composition (Cohort 1) –

On Day 2 post-injection, the cross-sectional area and aspect ratio of samples injected with any of the LPS doses were comparable to sham. On day 7, a dose of 1 or 10 μg/mL did not result in dimensional or biomechanical changes of motion segments. However, there was a significant increase (p<0.05) in cross sectional area from 10.92 mm² (SD: 0.73) in the sham group to 13.45 mm² (SD: 2.40) after injection with 100 μg/mL LPS, and a significant decrease (p<0.05) in aspect ratio to 2.69 mm/mm (SD: 0.36) after 100 μg/mL LPS injection from 3.40 mm/mm (SD: 0.39) in the sham group, which is represented as a dashed line (Figure 1A–B). Time dependent decreases in equilibrium modulus in the sham (p<0.05) and 10 μg/mL LPS (p<0.05) groups were observed at day 28 versus day 7 (Figure 1C–D). There were also non-significant trends of decreasing equilibrium (p=0.067) and dynamic moduli (p=0.057) in the 100 μg/mL LPS group compared to sham at day 7. Similar decreases in both mechanical properties (equilibrium modulus: p=0.082; dynamic modulus p=0.071) were observed at day 28 post-injection. There were no significant differences in mechanical properties between the control and sham groups (Supplemental Figure 1A–B).

Figure 1.

(A) Cross sectional area and (B) Aspect ratio of intervertebral discs after injection of LPS normalized to PBS sham indicated by the dashed line (n=2-6 per group). * p<0.05 LPS injection in comparison to normalized sham value indicated by dashed line. (C) Equilibrium modulus and (D) Dynamic modulus measurements for sham or LPS discs (n=4-18 per group). ^ p<0.05 day 28 in comparison to day 7.

In the AF, DNA content significantly increased (p<0.05) at day 7 post-injection of 10 μg/mL LPS, while other LPS doses had comparable DNA content to sham. The 1 μg/mL LPS injection caused a significant increase (p<0.01) in NP DNA content, with no significant differences observed at other LPS doses (Figure 2A). AF collagen content was similar to sham 7 days post-LPS injection. However, in the NP, a significant increase in collagen content was observed in the 100 μg/mL LPS group compared to sham (p<0.01, Figure 2B). At day 7, GAG content in the AF significantly increased (p<0.05) in the 100 μg/mL LPS group versus sham, but not in the 1 and 10 μg/mL LPS groups (Figure 2C). No significant differences in NP or AF GAG content were observed between sham and LPS or control and sham at day 28, although content levels decreased compared to those at day 7 in both the sham and LPS groups (Figure 2E, Supplemental Figure 1C). Water content of AF and NP for all groups at day 7 and day 28 showed no significant differences (Figure 2D, F). No significant difference in water content of the AF was observed between control and sham discs; however, the NP of the sham group had significant greater water content than the control (p<0.01, Supplemental Figure 2D).

Figure 2.

(A) DNA content, (B) Collagen content, (C) GAG content, and (D) Water content of PBS sham or LPS IVDs 7 days after injection (n=4-16 per group). (E) GAG content and (F) Water content of discs 28 days after injection (n=5-18 per group). * p<0.05 LPS injection in comparison to sham.

Behavior Testing: Mechanical and Thermal Sensitivity (Cohorts 2 & 3) –

The PAM test in the dorsal tail and von Frey tests in the ventral tail and hind paw were used to measure changes in mechanical sensitivity longitudinally over 28 days in comparison to baseline, and between the LPS and sham groups at each time point. In the incision only control group, the withdrawal force was significantly decreased at day 1 to 14 compared to baseline (Day 1-7: p<0.01; Day 14: p<0.05), but it recovered to baseline levels for day 21 and 28. A non-significant trend at day 21 was observed for the no incision control compared to sham (p=0.093, Supplemental Figure 2A). Withdrawal force significantly decreased compared to baseline from day 1 to 28 after injection in both the sham (Day 1: p<0.0001; Day 7: p<0.001; Day 14, 21: p<0.01; Day 28: p<0.05) and LPS (Day 1-28: p<0.0001) groups. However, the withdrawal force appeared to recover faster in the sham than the LPS group, resulting in a significantly lower PAM withdrawal force in the LPS group compared to sham at day 14 (p<0.05) and 28 (p<0.05) with a trend observed at day 21 (p=0.073, Figure 3A). In the von Frey test of the ventral tail, withdrawal force significantly decreased compared to baseline at day 1 through day 28 after LPS injection (Day 1-21: p<0.01; Day 28: p<0.05) with no significant decreases in the sham group (Figure 3B). Similarly, a significant decrease in withdrawal force compared to baseline was only observed at day 7 in the incision only control (p<0.01, Supplemental Figure 2B). In the hind paw, withdrawal force was similar to baseline up to day 7 in both groups. In the LPS group, the withdrawal force at day 7 was significantly decreased (p<0.05) compared to baseline, while in the sham group significance (p<0.0001) from baseline was observed at day 14. At day 21, the withdrawal force in the LPS group was significantly lower than sham (p<0.05, Figure 3C).

Figure 3.

(A-C) Behavioral tests measuring mechanical sensitivity including PAM (n=9-10 per group), Von Frey in the ventral tail (n=5 per group), and Von Frey in the hind paw (n=4-5 per group) at baseline and longitudinally 1 to 28 days after injection of saline sham or LPS IVDs. (D-E) Behavioral tests measuring thermal sensitivity including Tail Flick (n=9-10 per group) and Hargreaves (n=5 per group) at baseline and longitudinally 1 to 28 days after injection. * p<0.05 for LPS in comparison to sham, ^ p<0.05 LPS significantly different from baseline at indicated timepoint, + p<0.05 Sham significantly different from baseline at indicated timepoint.

The tail flick and Hargreaves behavioral tests were used to measure changes in thermal sensitivity after LPS injection. Neither thermal sensitivity test resulted in significant differences between the sham and LPS injection groups. In the tail flick test of the distal tail, significantly lower withdrawal time was measured 7 (p<0.05), 14 (p<0.01), and 28 (p<0.05) days after LPS injection compared to baseline. No differences were observed in the sham group versus baseline. In the incision only control, withdrawal time was significantly decreased compared to baseline from day 1 to 28 (Day 1: p<0.01; Day 7: p<0.05; Day 14: p<0.01; Day 21-28: p<0.05). In the sham group, the Hargreaves test of the ventral tail resulted in a significantly decreased withdrawal time from day 7 through day 28 (Day 7: p<0.01; Day 14, 21: p<0.001; Day 28: p<0.01) compared to baseline. Time dependent differences were not observed in the LPS group. However, a trend (p=0.078) towards higher withdrawal time in the sham group compared to LPS at baseline was observed. A significant difference was also observed at baseline between the sham and incision only control (p<0.05), as well as a decrease in withdrawal time for the control roup at day 14 compared to baseline (p<0.05, Supplemental Figure 2D).

Histology: Structural and Morphological Changes (Cohort 4) –

At days 14 and 28 post-injection, fluoroscopic images of all injected IVDs were analyzed and normalized to pre-injection images. DHI significantly decreased (p<0.001) in the LPS group over time from day 14 to 28. DHI in the LPS group was also significantly lower than sham at day 28 (p<0.01, Figure 4B). Representative images from each group and time point are shown to illustrate the heterogeneity of the histological changes (Figure 4A). Total histological score was significantly higher (p<0.05) in the LPS group compared to sham 28 days post-injection (Figure 4C). Of the 8 histological categories evaluated, 5 exhibited significantly greater scores between LPS and sham at day 28 including NP shape (p<0.01), NP area (p<0.05), NP-AF border appearance (p<0.01), AF lamellar organization (p<0.01), and AF tears/fissures/disruptions (p<0.05). A trend towards higher histological scores in the endplate of the LPS group was observed in comparison to sham (p=0.076). No significant differences in NP cell number and NP cell clustering and morphology were observed. Additionally, no significant differences were observed at day 14 in any histological grading category (Figures 4D–K).

Figure 4.

(A) Representative images of IVDs from saline sham and LPS groups at 14 and 28 days after injection. Scale bar = 1000 µm. (B) Change in disc height post-injection normalized to pre-injection height (n=12 per group). (C) Total histological score of sham and LPS-injected discs (n=6 per group). (D-K) Histological score categorical breakdown and score distributions (n=6 per group). * p<0.05 for indicated groups.

Gene Expression: Macrophage, Inflammatory, and Nociceptive Markers (Cohorts 4) –

AF and NP tissue regions were analyzed separately at day 14 and 28 for gene expression of markers related to macrophages, inflammatory cytokines, and nociception. Expression of the pro-inflammatory macrophage marker, Inos, was significantly increased in the AF after LPS injection at day 14 (p<0.01) and 28 (p<0.05). No significant changes in expression of the anti-inflammatory macrophage marker, Arg1, were found in the AF (Figure 5A). In the NP, Inos expression was also trending towards an increase in the LPS group compared to sham at day 14 (p=0.081). Expression of Arg1 was significantly decreased in the NP of LPS-injected discs compared to sham at day 14 (p<0.01); however, this trend reversed showing increased expression in the NP 28 days post-injection (p<0.01). There was also a significant increase (p<0.001) in NP Arg1 expression over time (Figure 5B). The pro-inflammatory cytokines, Tnfa and Il1b, were significantly increased in the AF of the LPS group in comparison to sham at day 28 (Tnfa: p<0.0001; Il1b: p<0.05), with a significant increase (p<0.0001) in Tnfa over time (Figure 5C). In the NP, an increase in Il1b was observed at day 14 in LPS-injected discs compared to sham (p<0.05), but no difference in Tnfa expression was observed. No changes in inflammatory cytokine expression were observed in the NP at day 28 (Figure 5D). Changes in the expression of nociceptive markers, Cgrp, Ngf, and Bdnf, were observed in the AF and NP at both 14 and 28 days post-injection. In the AF, Bdnf expression in the LPS group was trending higher at day 14 (p=0.097), while Cgrp and Ngf expression were similar to sham. At day 28, all three nociceptive markers were significantly increased in the LPS group in comparison to sham (Cgrp: p<0.0001; Ngf: p<0.05; Bdnf: p<0.001), and expression of Cgrp and Bdnf increased over time (Cgrp: p<0.0001; Bdnf: p<0.05, Figure 5E). In the NP at day 14, only Bdnf exhibited a trend towards increased expression in the LPS group compared to sham (p=0.064). Contrarily, Cgrp and Ngf were significantly increased at day 28 (Cgrp: p<0.01; Ngf: p<0.05, Figure 5F).

Figure 5.

Gene expression for macrophage markers in (A) AF and (B) NP tissue (n=6 per group). Expression of inflammatory genes in (C) AF and (D) NP tissue (n=6 per group). Expression of markers for nociception in (E) AF and (F) NP tissue (n=6 per group). * p<0.05 or indicated p-value LPS in comparison to saline sham, ^ p<0.05 Day 28 in comparison to day 14.

Immunostaining: Protein Expression of Macrophage and Nociceptive Markers –

Histological sections were analyzed for protein expression of macrophage and nociceptive markers. “No primary antibody” controls for each combination of antibodies confirmed antigen-specific staining for all markers (Supplemental Figure 4). The pan-macrophage marker CD68 was used to investigate the presence of macrophages, while colocalization with CCR7 or Arg1 was indicative of M1 or M2 macrophage subtypes, respectively (Figure 6A–B). Significant differences in CD68 expression were observed at Day 14 with increases observed after LPS injection compared to sham in the NP (p<0.05), EP+GP (p<0.05), and periphery (p<0.01) regions. At Day 28, CD68 expression only remained significantly increased in the NP (p<0.05, Figure 7A). Expression of the M1 marker, CCR7, was significantly increased in the LPS group at Day 14 in the periphery (p<0.0001) with a trend in in the EP+GP (p=0.062) region. At Day 28, the trend of increased CCR7 expression remained in the EP+GP (p=0.056) region while a trend for increased expression in the NP also emerged (p=0.054). In the periphery region, CCR7 expression significantly decrease over time after LPS injection (p<0.01, Figure 7B). Expression of the M2 marker, Arg1, was significantly increased in the NP 14 days after LPS injection compared to sham (p<0.05), with a non-significant trend of increased expression also at Day 28 (p=0.094). Interestingly, although no significant differences between sham and LPS were observed in the AF region, there was a significant increase in AF Arg1 expression over time between Day 14 and 28 (p<0.05, Figure 7C). Staining for the nociceptive marker PGP9.5 was used to demonstrate presence of all neurons in the IVD, while colocalization with CGRP represented nociceptive neurons. Representative images show PGP9.5 positive cells in both sham and LPS groups at Day 14 and 28, with most staining localized to the outer AF. 14 days after LPS injection, some cells positive for PGP9.5 exhibited colocalized staining with CGRP with a greater amount of colocalization at Day 28, while cells stained for PGP9.5 in the sham group were not CGRP positive (Figure 8).

Figure 6.

Representative images of (A) CD68 and CCR7 and (B) CD68 and Arg1 immunostained sections for the four quantified IVD regions 14 or 28 days after saline sham or LPS injection. Scale bar = 100 µm.

Figure 7.

MFI quantification of sham and LPS-injected IVDs stained for (A) CD68, (B) CCR7, and (C) Arg1 (n=6 per group). * p<0.05 for indicated groups, ^ p<0.05 Day 28 significantly different from Day 14 at indicated timepoint.

Figure 8.

Representative images in the outer AF of PGP9.5 and CGRP immunostained sham and LPS-injected IVDs 14 and 28 days after injury. Arrow indicates PGP9.5 or CGRP staining. Scale bar = 100 µm.

Serum Analysis: Systemic Cytokine Changes –

Blood serum was analyzed for changes in cytokine concentrations longitudinally over 28 days by normalization to baseline levels, represented as a dashed line. Pro-inflammatory cytokines, TNFα, IL-1β, IL-1α, interferon gamma (IFNγ), IL-5, IL-6, IL-12p70, and IL-17A all exhibited higher levels initially at day 7 in the LPS group compared to baseline, followed by an increase of greater magnitude 28 days post-injury. This was demonstrated by a significant increase in TNFα (p<0.05; Sham: 442.4±361.8 ng/mL; LPS: 1443±1024 ng/mL), IFNγ (p<0.05; Sham: 579.7±472.3 ng/mL; LPS: 1781±1410 ng/mL), IL-6 (p<0.05; Sham: 542.6±431.6 ng/mL; LPS: 1263±1121 ng/mL), IL-12p70 (p<0.05; Sham: 1196±803.1 ng/mL; LPS: 2465±1724 ng/mL), and IL-17A (p<0.05; Sham: 102.4±53.29 ng/mL; LPS: 125±96.59 ng/mL) 7 days after LPS injection compared to sham, with a trend in IL-5 (p=0.069; Sham: 929.7±429.5 ng/mL; LPS: 1111±531.9 ng/mL). IFNγ levels in the sham group were also significantly decreased compared to baseline at day 7 (p<0.05; Baseline: 1219±522.3 ng/mL; Day 7: 579.7±472.3 ng/mL). At day 21, IL-1α (p<0.05; Sham: 419.7±294.5 ng/mL; LPS: 460.2±310.9 ng/mL) and IL-1β (p<0.05; Sham: 259.7±206.1 ng/mL; LPS: 287.1±187.4 ng/mL) were significantly increased after LPS injection. Finally, significant increases between sham and LPS were observed in IL-1β (p<0.05; Sham: 715.7±457.2 ng/mL; LPS: 884.5±516.7 ng/mL), IL-1α (p<0.05; Sham: 1216±756 ng/mL; LPS: 1636±997.6), IL-5 (p<0.05; Sham: 1466±547.1 ng/mL; LPS: 1591±712.7 ng/mL), and IL-6 (p<0.05; Sham: 2234±1783 ng/mL; LPS: 3630±2369 ng/mL) at day 28 with a trend in IL-17A (p=0.055; Sham: 272.4±169 ng/mL; LPS: 323.8±207.8 ng/mL). Levels of all pro-inflammatory cytokines except TNFα were also significantly increased 28 days after LPS injection compared to baseline, including IL-6 (p<0.001; Baseline: 710.3±474.5 ng/mL; Day 28: 3630±2369 ng/mL), IL-5 (p<0.01; Baseline: 731.9±371.2 ng/mL; Day 28: 1591±712.7 ng/mL), IL-1β (p<0.01; Baseline: 202.8±122.4 ng/mL; Day 28: 884.5±516.7 ng/mL), IL-1α (p<0.01; Baseline: 262.3±179.9 ng/mL; Day 28: 1636±997.6 ng/mL), IFNγ (p<0.01; Baseline: 981.6±665.7 ng/mL; Day 28: 3214±2094 ng/mL), IL-12p70 (p<0.01; Baseline: 1035±681.0 ng/mL; Day 28: 3765±2347 ng/mL), and IL-17A (p<0.05; Baseline: 63.97±50.36 ng/mL; Day 28: 323.8±207.8 ng/mL, Figure 9A–H). The anti-inflammatory cytokine IL-4 was significantly increased 7 days post-LPS injection compared to sham (p<0.05; Sham: 328.0±206.8 ng/mL; LPS: 496.0±370.0 ng/mL); however, the difference did not remain significant at later timepoints (Figure 9I). Macrophage promoting cytokines, macrophage colony-stimulating factor (M-CSF), granulocyte colony stimulating factor (G-CSF), and macrophage inflammatory protein-1 alpha (MIP-1α), were also measured. M-CSF was significantly increased at day 7 (p<0.05; Sham: 73.51±64.66 ng/mL; LPS: 88.00±67.08 ng/mL), day 21 (p<0.05; Sham: 49.20±46.27 ng/mL; LPS: 53.86±44.43 ng/mL), and day 28 (p<0.01; Sham: 232.4±170.1 ng/mL; LPS: 372.0±240.0 ng/mL) in the LPS group compared to sham, and at day 7 (p<0.05; Baseline: 33.13±23.18 ng/mL; Day 7: 88.00±67.08 ng/mL) and day 28 (p<0.001; Baseline: 33.13±23.18 ng/mL; Day 28: 372.0±240.0 ng/mL) compared to baseline. Significant increases were seen in G-CSF at 7 days (p<0.05; Sham: 7.302±4.952 ng/mL; LPS: 21.13±14.51 ng/mL) and 28 days (p<0.05; Sham: 41.89±32.86 ng/mL; LPS: 73.57±50.87 ng/mL) after LPS injection. Interestingly, a decrease compared to baseline was observed in the sham group at day 7 (p<0.05; Baseline: 15.13±4.715 ng/mL; Day 7: 7.302±4.952 ng/mL) and the LPS group at day 14 (p<0.01; Baseline: 9.244±6.903 ng/mL; Day 14: 3.644±2.865 ng/mL), while G-CSF increased in the LPS group at day 28 (p<0.05; Baseline: 9.244±6.903 ng/mL; LPS: 73.57±50.87 ng/mL). MIP-1α was significantly increased in LPS compared to sham at Day 21 (p<0.05; Sham: 48.50±22.62 ng/mL; LPS: 46.95±25.14 ng/mL) and 28 (p<0.05; Sham: 139.7±81.89 ng/mL; LPS: 161.9±82.44 ng/mL) with an increase compared to baseline 28 days post-injury (p<0.01; Baseline: 35.27±18.41 ng/mL; Day 28: 161.9±82.44 ng/mL). Other cytokines were also measured but did not result in any statistically significant between group differences (Supplemental Figure 5A–K).

Figure 9.

(A-L) Blood serum cytokine levels measured longitudinally at baseline and 7 to 28 days after injection of saline sham or LPS IVDs (n=3-5 per group). * p<0.05 for LPS in comparison to sham, ^ p<0.05 LPS significantly different from baseline at indicated timepoint, + p<0.05 Sham significantly different from baseline at indicated timepoint.

Discussion:

This study focused on the development of a disc degeneration model in the rat caudal spine induced by inflammatory activation of the IVD using LPS injection into the NP space. Local injection of LPS triggered an inflammatory response indicated by increased gene and protein expression of markers for macrophages and pro-inflammatory cytokines locally in the IVD, as well as systemic changes in blood cytokine and chemokine levels. The inflammatory cascade caused structural changes in the disc, as well as biochemical alterations suggestive of ECM remodeling. Rats also exhibited greater local mechanical hyperalgesia in the spinal region of injury after LPS injection, a corresponding increase in nociceptive marker gene expression, and evidence of neuron ingrowth and sensitization in the injured IVD.

LPS Activation of TLRs Promotes Inflammatory Signaling –

Recent studies have focused on the activation of TLRs during disc degeneration and their role in propagating the production of inflammatory cytokines. LPS is a well-established agonist of TLR4, which is expressed on the surface of IVD cells and increases with higher degeneration grades²⁶. When stimulated with LPS in vitro, IVD cells have shown increased gene expression levels and tissue protein levels for TLR4 and the inflammatory cytokines, TNFα and IL-1β^27,28,32. This study extends these finding in vivo, where the response of IVD cells to LPS injection increased tissue gene expression of Tnfa and Il1b (Figure 5C–D). Importantly, the increased expression of inflammatory cytokines was sustained for 28 days post-injury and led to long-term, systemic increases of pro-inflammatory cytokines in the blood, although the location of increased inflammatory gene expression in the disc and concentrations of serum cytokines varied over time (Figure 9A–H, Supplemental Figure 5A–C). Similarly variable timelines for protein production of the inflammatory cytokines TNFα and IL-1β were demonstrated by Rajan et al. They showed an overall increase in inflammatory cytokine levels 1 day after in vivo LPS injection into the caudal spine, followed by a decrease in TNFα at day 7 while IL-1β continued to increase²⁸. Further corroborating these results, Ponnappan et al. observed increased Il1b gene expression in the NP, but not in the AF, 10 days after inflammatory stimulation in an explant model⁴⁴. Maintenance of long-term cytokine expression and protein production caused by caudal intradiscal LPS injection validates the model in the current study, as it mimics the presence of chronic inflammatory activation seen in human disc degeneration.

Other injury models have attempted to recreate this sustained inflammatory environment. While traditional puncture or lesion injury has been used previously to activate cytokine production in the IVD, this method does not result in a sustained inflammatory profile and causes physical trauma to the tissue that is not representative of native disc degeneration^16,20. To prevent the inflammatory resolution seen in single puncture or lesion injuries, some studies have utilized multiple injuries to the same disc or injection of inflammatory cytokines, such as TNFα, during puncture. While maintenance of pro-inflammatory cytokine expression and production is improved in these models, the tissue disruption from physical injury is amplified^18,22,23,45. In contrast, the use of a small 33G needle in this study prevents severe structural trauma to the IVD, while still inducing sustained pro-inflammatory effects from LPS injection.

Inflammatory Environment Leads to Extracellular Matrix Disruption –

Previous studies have demonstrated disruption to matrix integrity via decreased disc height and increased histological score using puncture injury or mechanical overloading. These models mimic the stimuli that disrupt the structural integrity of the IVD in human disc degeneration^{14,16,18,19,46}. This study expands on these findings, showing moderate degenerative changes to the structural integrity of the disc stimulated by exposure to the inflammatory stimulus, LPS, rather than overt physical trauma to the tissue. At 28 days post-LPS injection, a significant decrease in disc height and an increase in total histological score was observed (Figure 4A–C). Interestingly, when histological score is analyzed by subcategory, it suggests that the majority of degenerative changes occurred in the ECM of both the AF and NP, while the endplate and IVD cells appear largely unaffected (Figure 4D–K).

Extracellular matrix disruption is a common measure of degeneration severity, with decreasing ECM content generally correlating with increasing degeneration in human tissue samples³. Additionally, human NP and AF cells treated with an inflammatory stimulus have been shown to express decreased collagen and aggrecan^5,47. These studies support the role that pro-inflammatory cytokines play in the process of ECM degradation during degeneration. This has been further investigated in vivo, showing increased aggrecanase, an enzyme involved in the breakdown of the ECM component aggrecan, after injection of LPS or IL-1 into rat caudal IVDs²⁸. Biomechanical testing results indicated a trend towards loss of compressive mechanical properties after LPS injection, suggestive of the softer material that would result from ECM catabolism (Figure 1C–D, Supplemental Figure 1A–B). However, IVDs in this study had minimal changes in GAG and water content and an increase in collagen content of the NP after LPS injection (Figure 2B–F). Although there was an increase in the water content of the NP after saline injection compared to incision only controls suggesting a hydrating effect from the injections, the overall biochemical content findings, specifically the increase in collagen content, and the histological changes could indicate that LPS stimulation is having an effect on the ECM composition of the IVD, which is consistent with what is observed during human disc degeneration (Supplemental Figure 1C–D)⁴⁸.

Cytokines Encourage Immune Cell Recruitment –

In the healthy state, the IVD is considered an immune-privileged organ due to a lack of blood vessels and nerves in the majority of the tissue, with the exception of the outer AF. When the structural integrity is compromised, blood vessels grow into the disc from adjacent vertebral bodies or through damaged AF⁴. A trend towards increasing levels of VEGF, an angiogenic factor, in the blood of LPS injected rats was observed possibly supporting this mechanism, although this was not directly examined in this study (Supplemental Figure 5K)⁴⁹. Tissue damage and angiogenesis provide an opportunity for macrophages and other immune cells to infiltrate the disc, which is further enhanced by production of cytokines and chemokines by NP cells^47,50,51.

Macrophages can be broadly categorized as either M1, pro-inflammatory macrophages, or M2, anti-inflammatory macrophages. Each subtype plays a specific role in the degenerative environment, motivating them to be present in the IVD at differing times during the degenerative cascade. This has been shown previously in an in vivo injury model using puncture of the mouse caudal IVD 10 times. They observed increased expression of M1 macrophage markers from day 1 to 14, followed by a delayed increase in M2 macrophage markers from day 7 to 28 post-injury⁴⁵. The current study agrees with these findings, evidenced by a greater initial increase in gene expression of the M1 marker, Inos, at day 14 followed by an increase in the M2 marker, Arg1, in the NP at day 28, although Inos remained elevated in the AF (Figure 5A–B). Similarly, regional analysis of protein expression yielded widespread increases in CD68 positive macrophage staining at day 14, with less expression exhibited at day 28. Regional changes were also observed over time, with a decrease in the M1 marker, CCR7, in the periphery and an increase in the M2 marker, Arg1, in the AF over time (Figure 6A–B, 7A–C). Interestingly, the periphery region is rich in immune cells and may represent an injury capsule or scar tissue. Chemokines and factors that promote macrophage survival were also increased systemically up to 28 days after LPS injection (Figure 9J–L, Supplemental Figure 5D–H). However, only one anti-inflammatory cytokine, IL-4, was significantly increased after LPS injection and levels were not sustained over time, indicating that the microenvironment stimulated by LPS injection may promote pro-inflammatory M1 macrophages more than an M2 anti-inflammatory macrophage phenotype (Figure 9I, Supplemental Figure 5I–J). Further, other studies have shown that macrophage depletion mitigates the increase in tissue expression of inflammatory cytokines, Tnfa and Il1b, that occurs after injury⁵². This supports the role of macrophages in the production of cytokines and perpetuating the pro-inflammatory environment during disc degeneration⁵¹.

Pain Behavior Stimulated by the Degenerative Environment –

One mechanism for the pain associated with disc degeneration is from pathological ingrowth of neurons into the disc. It has been suggested that nerve cells enter the disc due to loss of IVD ECM which normally acts to inhibit neuronal ingrowth⁵⁰. However, symptomatic disc degeneration is comprised of two important components: nerve cell ingrowth into the IVD and sensitization of those neurons⁵³. Previous studies have looked at the progression of innervation after injury and how this manifests as pain behavior. It has been shown that local mechanical hyperalgesia increases in the acute phase of injury¹⁷. This was also observed in the current study where decreased force after an incision was observed in the PAM and von Frey tests; however, a more severe response was elicited after LPS injection compared to the sham group without the inflammatory stimulus (Figure 3A–B, Supplemental Figure 2A–B). Prior studies of injury to the lumbar IVD found that radiating mechanical sensitivity in the hind limb increased later, at 3-9 months after puncture injury in a mouse and at least 4 weeks after injury in a rat^14,15. The different timeframe for observing a radiating pain phenotype could explain why greater differences in pain behavior were observed local to the site of injury in the current study, although the difference in spinal level and type of injury may also have contributed (Figure 3C). Additionally, it has been hypothesized that early thermal sensitivity occurs from exposure of neurons to NP contents released during injury⁵⁴. Indeed, Mohd Isa et al. showed increased thermal hyperalgesia 2 to 29 days after puncture injury¹⁷. Contrarily, other studies have demonstrated a shorter initial response to thermal stimuli; however, it reappeared 9-12 months after puncture injury, with most innervation identified via staining in the dorsal region prior to 6 months post-injury^14,50. Similarly, our results showed limited responsiveness to thermal stimuli up to 28 days after LPS injection (Figure 3D–E, Supplemental Figure 2C–D). A longer timeframe or additional testing sites could potentially be utilized to further assess the inflammatory effects of LPS injection on thermal hyperalgesia.

Nonetheless, inflammatory cytokines are known to play a critical role in the induction of pain, with greater cytokine levels being correlated with higher degrees of pain^55,56. Indeed, an injection of the cytokine TNFα was shown to increase degeneration grade and promote a pain phenotype in comparison to no injection. However, when a TNFα antagonist was injected instead, both degenerative and pain effects were mitigated²⁵. It has also been demonstrated that treatment of IVD cells with TNFα or IL-1β stimulates production of NGF, a regulatory protein for nerve growth⁴⁵. Expanding these results in vivo, this study demonstrated a similar increase in tissue gene expression of Ngf after inflammatory activation using LPS injection. Additionally, CGRP is known to be involved in the sensitization of neurons with increased expression being shown after injury⁵⁷. The results of this study confirm this finding, demonstrating increased gene expression of Cgrp at day 28 in both the NP and AF (Figure 5E–F). Nerve cell ingrowth stimulated by NGF and sensitized by CGRP provides a possible mechanism for the pain behaviors demonstrated in this inflammatory injury model. Protein expression of PGP9.5 in the outer AF in both the sham and LPS groups indicates the presence of neurons in both groups, with increased CGRP colocalization only in the LPS injection group demonstrating nociceptive sensitization as a possible pain-mediating mechanism after LPS injection (Figure 8).

Limitations –

The limitations of this study include the timeframe of 28 days and the use of only male rats. The chosen timeframe of 28 days is an extension of those used in some previous studies; however, in long-term models over a period of 12 months, degeneration grade continued to change over time and some pain behaviors reemerged at later time points^14,50. Moreover, the recovery from an incision in all groups was a complicating factor in the interpretation of pain behavior over this timeframe, and may be alleviated in future studies using percutaneous LPS injection as an alternative method of delivery. The use of only male rats limits the generalization of the findings because sex-specific pain behaviors and IVD biomechanics have been identified. If female rats had also been used, pain behavior results may have been more variable based on a study by Mosley et al. In injured male rats, they observed significant pain responses compared to control, while female rats exhibited large variability in both sham and injury groups resulting in no significant differences in pain behavior¹⁸.

Conclusion –

This study demonstrates a persistent local and systemic inflammatory response, structural changes, and pain behavior as a result of intradiscal inflammatory stimulation of the IVD in vivo. Together, these findings indicate that inflammation independent of traumatic IVD structural damage, triggers the degenerative cascade. Future studies will utilize this rat injury model to delineate the mechanism at play in the relationship between inflammation and discogenic pain during disc degeneration.

Data Availability Statement:

The data that support the findings of this study are presented in the results and supplemental material of this article.