Proinflammatory intervertebral disc cell and organ culture models induced by tumor necrosis factor alpha

Abstract Inflammation plays an important role in the pathogenesis of intervertebral disc (IVD) degeneration. The proinflammatory cytokine tumor necrosis factor alpha (TNF‐α) has shown markedly higher expression in degenerated human disc tissue compared with healthy controls. Anti‐inflammatory treatment targeting TNF‐α has shown to alleviate discogenic pain in patients with low back pain. Therefore, in vitro and ex vivo inflammatory models utilizing TNF‐α provide relevant experimental conditions for drug development in disc degeneration research. The current method article addressed several specific questions related to the model establishment. (a) The effects of bovine and human recombinant TNF‐α on bovine nucleus pulposus (NP) cells were compared. (b) The required dose for an inflammatory IVD organ culture model with intradiscal TNF‐α injection was studied. (c) The effect of TNF‐α blocking at different stages of inflammation was evaluated. Outcomes revealed that bovine and human recombinant TNF‐α induced equivalent inflammatory effects in bovine NP cells. A bovine whole IVD inflammatory model was established by intradiscal injection of 100 ng TNF‐α/ cm3 disc volume, as indicated by increased nitric oxide, glycosaminoglycan, interleukin 6 (IL‐6), and interleukin 8 (IL‐8) release in culture media, and upregulation of MMP3, ADAMTS4, IL‐8, IL‐6, and cyclooxygenase (COX)‐2 expression in NP tissue. However, results in human NP cells showed that the time point of anti‐inflammatory treatment was crucial to achieve significant effects. Furthermore, anticatabolic therapy in conjunction with TNF‐α inhibition would be required to slow down the pathologic cascade of disc degeneration.

The current method article addressed several specific questions related to the model establishment. Outcomes revealed that bovine and human recombinant TNF-α induced equivalent inflammatory effects in bovine NP cells. A bovine whole IVD inflammatory model was established by intradiscal injection of 100 ng TNF-α/ cm 3 disc volume, as indicated by increased nitric oxide, glycosaminoglycan, interleukin 6 (IL-6), and interleukin 8  release in culture media, and upregulation of MMP3, ADAMTS4, IL-8, IL-6, and cyclooxygenase (COX)-2 expression in NP tissue. However, results in human NP cells showed that the time point of anti-inflammatory treatment was crucial to achieve significant effects. Furthermore, anticatabolic therapy in conjunction with TNF-α inhibition would be required to slow down the pathologic cascade of disc degeneration.

K E Y W O R D S
3R, cytokines, inflammation, intervertebral disc, regeneration, spine

| INTRODUCTION
Low back pain (LBP) is the leading cause of disability worldwide. 1 One major cause for chronic LBP is symptomatic intervertebral disc degeneration (IVDD). [2][3][4] IVDD is characterized by extracellular matrix (ECM) degradation, accelerated cartilaginous, and bone remodeling, release of proinflammatory cytokines, altered spine biomechanics, angiogenesis and neoinnervation, altogether potentially leading to chronic LBP, and disability. [5][6][7][8][9] IVDD can be induced by mechanical stress, trauma, infection, genetic predisposition, and inflammation. [10][11][12][13][14][15][16] Inflammation plays a major role in disc degeneration, as proinflammatory cytokines (ie, tumor necrosis factor alpha [TNF-α], interleukin 1 beta [IL-1β], interleukin 6 [IL-6], interleukin 8 , interleukin 17 , and interferon gamma [IFN-γ]) induce and trigger discal ECM breakdown and accelerated catabolism by stimulation of catabolic enzymes such as matrix metalloproteinases (MMPs) and a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS). 6,[17][18][19][20] Proinflammatory cytokines have shown elevated expression in degenerative and symptomatic compared to healthy and asymptomatic IVDs. 16,21 Since therapeutic approaches for IVDD remain limited, biological anti-inflammatory approaches to IVD regeneration have gained increasing interest. In cases of refractory LBP due to IVDD, anti-inflammatory and/or anti-degenerative therapies such as cytokine inhibition may relieve pain and slow down the progression of the disease. [22][23][24][25][26][27][28][29][30][31] Several studies indicated that cyclooxygenase-2 (COX2) inhibitors can reduce the inflammatory response in different models. [25][26][27] Soluble TNF receptor type II is able to significantly attenuate the effects of TNF-α on primary human IVD cells in vitro. 28 Intradiscal administration of a TNF-α inhibitor, Etanercept, in LBP patient can alleviate intractable discogenic LBP for up to 4 weeks. 31 A degenerative disc exhibits increased TNF-α expression, not only produced by immunocytes, but also by disc cells themselves. 15,19,32 Furthermore, TNF-α can induce nucleus pulposus (NP) cells to produce other cytokines and chemokines that can further enhance the inflammatory state by recruiting and activating immune cells. 33 So far, it is widely accepted that TNF-α contributes to disc degeneration by decreasing the anabolism and increasing the catabolism of ECM. 34 Additionally, exogenous TNF-α induces neuropathology and sensory nerve growth into IVD, which indicated TNF-α might be the chemical mediator of discogenic pain. 35,36 Therefore, multiple in vitro, ex vivo, and in vivo inflammatory IVD models have been established with TNF-α. 20,28,[37][38][39] NP cells cultured with TNF-α in vitro showed   upregulated expression of catabolic enzymes, ADAMTS 4&5 and   MMP-1,-2,-3,-13, and inflammatory mediators, IL-1β, IL-6, IL-8, and   COX2, downregulated expression of ECM markers collagen II, aggrecan, and versican. 34,40-44 TNF-α has been shown to induce MMP3 expression via nuclear factor κB (NF-κB) and mitogen-activated protein kinase pathways. 45 Intradiscal injection of TNF-α in a porcine model was sufficient to induce early-stage disc degeneration, characterized by matrix loss, annular fissure formation, and vascularization. 46 Lai et al reported that annular puncture with TNF-α injection enhanced painful behavior with disc degeneration in a rat model. 39 Ex vivo explant culture models bridge the gap between in vitro and in vivo systems and reveal many advantages by maintaining the native tissue environment and decreasing the consumption of experimental animals. Compared with the small animals like mouse, rat and rabbit, the IVDs from large animals such as sheep, dog and cow are more similar to human. They show comparable size and loss of notochordal cells in early adulthood as human IVD. 47,48 Notochordal cells have been reported to present anti-inflammation and regenerative effect in IVDs. 49,50 With those similarities, many bovine caudal IVD organ culture models were established. van Dijk et al developed a NP tissue explant culture model, and found that using polyethylene glycol to raise culture medium osmolarity was able to maintain the NP tissue specific matrix composition. 25,51 Whole bovine caudal IVD cultured under   either limited glucose condition or high-frequency loading condition led to   a significant drop in cell viability, while combined treatment with limited glucose and high-frequency loading resulted in an additive increase in cell death in both the NP and annulus fibrosus (AF), and an increase in MMP13 gene expression. 52 Purmessur et al cultured whole IVD organ excluding the endplates with exogenous TNF-α in medium. Aggrecan degradation products and β-galactosidase staining were enhanced by TNF-α on day 21 without any recovery, when TNF-α was removed on day 7. 38 Recently, our group has developed a proinflammatory and degenerative IVD whole organ culture system to investigate the proinflammatory and degenerative microenvironment operant in IVDD. Results indicated that a combination of detrimental dynamic loading, nutrient deficiency and intradiscal TNF-α injection could synergistically simulate the proinflammatory and degenerative disease condition. However, intradiscal TNF-α injection alone did not lead to a significant inflammatory effect. 7 In the present study, we sought to establish TNF-α induced in vitro and ex vivo IVD inflammation models, which would represent preclinical testing systems for screening of anti-inflammatory drugs for disc degeneration treatment. Specifically, the following questions were addressed within this study:  53 In the current study, human and bovine NP cells isolation and expansion were performed with αMEM according to previous publication. 54 DMEM contains much higher amount of vitamins, amino acids and glucose than αMEM. Therefore, cells and IVD organ culture experiments with TNF-α and Etanercept were performed with DMEM, due to a much higher cell density and nutrition requirement in these experiments.

| NP cells isolation and expansion
Human NP cells were isolated from traumatic IVDs (2 donors, 34/ 49 years old, male) with ethical approval (Cantonal Ethic Commission Bern 2016). General consent was obtained from all patients before surgery. All studies were performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. The IVDs were classified as mildly degenerated by MRI (Pfirrmann grade 2-3). Bovine NP cells were isolated from caudal intervertebral discs of 6 to 12month-old calves from local abattoirs immediately after death. NP cell isolation was performed as described previously. 55 The collected NP tissue was cut into small pieces. Human NP tissue was incubated with

| Effect of human and bovine recombinant TNF-α on bovine NP cells
Bovine NP cells were seeded at a concentration of 60 000/cm 2 in 12well plates with DMEM medium (containing 4.5 g/L glucose) supplemented with 10% FBS. After cell attachment (24 hours after cell seeding), the medium was exchanged to serum-free experimental medium (DMEM supplemented with 1% ITS+, 1% nonessential amino acid [NEAA, Gibco, Paisley, UK], 50 μg/mL ascorbate 2 phosphate and 1% P/S) with or without inflammatory inducers 10 ng/mL human recombinant TNF-α (R&D systems, Zug, Switzerland) or 10 ng/mL bovine recombinant TNF-α (R&D Systems, Zug, Switzerland). After another 72 hours of culture, the cell monolayer was lysed and RNA was isolated for gene expression analysis.

| Effect of human recombinant TNF-α and TNF-α inhibition on human NP cells
Human NP cells were seeded into a six well-plate at a cell density of 30 000/cm 2 . One day after seeding, cells were treated with 10 ng/mL (low dose) or 50 ng/mL (high dose) TNF-α in serum-free experimental medium as described above for bovine NP cells TNF-α experiments.
The samples were collected at three timepoints, 6, 24, and 48 hours after treatment, for gene expression analysis.
To investigate the effect of TNF-α blocking with the TNF-α inhibitor Etanercept (Enbrel, Pfizer, New York, New York), NP cells were seeded as described above and cultured for 24 hours to allow for cell attachment. Hereafter, cells were divided into 4 different groups: (1) iNP-cells were treated with 10 ng/mL TNF-α for 48 hours, (2) iNP-Eta-cells were treated with 10 ng/mL TNF-α and immediately after 1 μg/mL Etanercept was added for 48 hours, (3) iNP-24 hours-Eta-cells were treated with 10 ng/mL TNF-α, 24 hours after 1 μg/mL Etanercept was added, and (4) iNP-24 hours-FM-cells were treated with 10 ng/mL TNF-α, 24 hours after replaced to fresh medium without TNF-α. Cells treated with serum-free culture medium as described above served as negative control. All the cells were harvested for gene expression analysis at 72 hours after seeding. The concentration of Etanercept used here was selected according to previous studies, showing that Etanercept at 0.01, 0.1 and 1 μg/mL induced less than 8% cell death in TNF-α transfected Jurkat cells, and in human NP cells and AF cells cultured with Etanercept at 100, 250, 500, 1000, and 2000 μg/mL, cell proliferation was only suppressed with Etanercept at 500 μg/mL or higher. 56,57 Therefore, the selected Etanercept concentration at 1 μg/mL was assumed to have no cytotoxic effect on NP cell culture in vitro.

| IVDs dissection
Bovine caudal IVDs were collected from fresh sacrificed 6 to 12month-old calves from local slaughterhouses. Disc dissection was performed as described previously. 58 Briefly, most of the muscle and soft tissue were removed, whole IVDs with cartilage endplates (EPs) were isolated with a band saw and redundant vertebral bone and growth plate were carefully cut off to ensure two parallel planes of discs. Disc height and diameter was then measured with a caliper. Disc volume = (long diameter + short diameter)/2) 2 × π × disc height. The surfaces of EPs were cleaned using a Pulsavac Wound Debridement Irrigation System (Zimmer, Minneapolis, USA) with Ringer's buffer to remove the cutting debris and blood clots. After prewashing in PBS with 10% P/S, IVDs were cultured in 6-well plates with 7.5 mL IVD culture medium, DMEM supplemented with 1% P/S, 50 mg/mL Primocin (Invitrogen, San Diego, California), 2% FBS, 50 μg/mL ascorbate 2 phosphate, 1% ITS+, 1% NEAA, at 37 C, 5% CO 2 .

| IVD culture and intradiscal injection
IVDs having a diameter of 1.5 to 2.0 cm were selected for the current study. IVDs were cultured free swelling during the night. Dynamic loading was performed, at 0.02 to 0.2 MPa, 0.2 Hz for 2 hours per day within a bioreactor. 7 IVDs from each donor were randomly divided into three groups: PBS, TNF-α and TNF-α + Etanercept. TNFα + Etanercept: 40 μL of TNF-α, containing 100 ng TNF-α/cm 3 of disc volume, was firstly injected into the disc, 30 minutes after 20 μL Etanercept, containing 10 μg Etanercept per 100 ng TNF-α, was injected into the disc. TNF-α: 40 μL of TNF-α, containing 100 ng TNFα/cm 3 of disc volume, was injected into disc 30 minutes after 20 μL PBS was injected. PBS: 40 and 20 μL PBS was injected into disc sequentially. The injection was performed using a 30-gauge insulin needle, after the first dynamic loading on day 1. The intradiscal injection dose of Etanercept was kept at the same ratio of TNFα to Etanercept as in vitro, which is 1:100. IVDs were cultured with daily dynamic loading and free swelling recovery overnight, the disc size of IVDs was measured before and after loading. 7 Culture media were collected daily after free swelling for further analysis. The NP tissue (gel-like inner core of 6-8 mm of diameter) was collected for gene expression analysis on day 2 and day 5.

| Glycosaminoglycan and nitric oxide measurement
The amount of Glycosaminoglycans (GAGs) released in IVDs culture media was measured by using the 1,9-dimethylmethylene blue dye method. 60 The level of GAG release from each IVD at each time point after injection was normalized to the amount released on day 1 before injection by dividing the corresponding day's GAG release content with the amount of GAG release on day 1. The concentration of nitric oxide (NO) in the culture media of IVDs was detected as the level of its stable oxidation product, nitrite (NO 2− ), using the Griess Reagent Kit (Promega, USA). The NO concentrations in the media are presented in the results section without normalization.

| Statistical analysis
Statistical analyses were performed using the GraphPad Prism 7 software (GraphPad Software, Inc., La Jolla, California). D'Agostino-Pearson omnibus normality test was used to define whether the data were normally distributed. For data that were normally distributed, unpaired t-test was used to determine differences between two groups; One-way ANOVA was used to determine differences between three or more groups. For the not normally distributed data, Mann-Whitney U test was used to determine differences between two groups; Kruskal-Wallis test was used to determine differences between three or more groups. P < .05 was considered statistically significant.

| Proinflammatory IVD organ culture model
According to our previous study, intradiscal injection of 100 ng human TNF-α per disc did not induce a significant inflammatory effect. 7 As shown in Figure 1, (0.72-1.49)) expression were not changed by TNF-α. After free swelling disc height increased by proximately 5% and disc volume increased by proximately 18%. After daily loading disc height decreased by approximately 10% and disc volume by approximately 5%, compared with day 0 when discs were isolated. However, the fold changes of disc height and volume did not show any difference among these three groups ( Figure 6).

| DISCUSSION
Anti-inflammatory therapy has been considered as a promising approach to delay the IVD degeneration and relieve discogenic pain.
TNF-α, as a pro-inflammatory factor, has been reported to be associated with IVD degeneration and discogenic pain. 21,61 Anti-inflammatory therapies targeting TNF-α are widely reported, with preserved matrix production and restraint of matrix degradation. 31,37,42 Therefore, in vitro and ex vivo IVD inflammatory culture systems induced by TNF-α are clinically relevant models for drug development for treatment of disc degeneration.
In the current study, several specific questions related to the inflammatory model were investigated. Firstly, due to the scarce access to human IVD tissue and especially to healthy samples, bovine IVD cells and bovine caudal whole IVDs have been widely used in spine research. While using TNF-α for inflammation induction of bovine disc cells or organs, one question which has not been well addressed is whether human and bovine recombinant TNF-α imply the same effect on bovine disc cells, or whether the TNF-α receptors on bovine disc cells can also transmit the signaling from human F I G U R E 3 Relative mRNA expression levels of human NP cells. Human NP cells were cultured with basal medium (Control), TNF-α 10 ng/mL for 48 hours (iNP), TNF-α immediately followed by 1 μg/mL Etanercept for 48 hours (iNP-Eta), TNF-α for 24 hours followed by 1 μg/mL Etanercept added for 24 hours (iNP-24 hours-Eta), or TNF-α for 24 hours followed by fresh basal medium without TNF-α Our result showed that TNF-α downregulated gene expression of type II collagen in cell culture, but not in whole organ culture. In contrast, Seguin et al showed that TNF-α decreased expression of type II collagen in NP tissue culture. 34 This suggests that the whole IVD organ culture system is beneficial for maintaining disc cell homeostasis, which may be due to the physiological osmolarity inside the intact organ that has been shown to maintain the NP tissue specific matrix composition. 51 Annular puncture may induce disc degeneration depending on the disc size and needle size. 63 A recent goat study revealed that 22G needle puncture did not result in degenerative changes in lumbar IVDs, nor was degeneration found in IVDs of Beagles injected using 25G needles. 64,65 Also in bovine caudal IVD we found that IVD puncture using a 30-gauge needle did not cause dysregulation on expression of anabolic, catabolic and inflammatory markers. 7 Therefore, injection using a 30-gauge needle is not expected to cause an effect on the state of IVD degeneration in the current experiments.
F I G U R E 4 NO, IL-6, IL-8, and GAG release in the IVD culture medium. NO (A), IL-6 (B), IL-8 (C), and relative GAG (D, normalized to day 1) release in the conditioned medium of IVDs with PBS injection (PBS), TNF-α injection (TNF-α), and TNF-α plus Etanercept injection (TNFα + Etanercept). Intradiscal injection performed after day 1 loading. Mean + SD, n = 9, *P < .05, **P < .01, and ***P < .001 Analysis of the culture medium was undertaken to investigate whether the molecule release was related to the disc volume. This was performed with IVDs cultured under physiological loading and without TNF-α injection. The initial GAG release on day 1 from discs of different donors showed a high variation, which may result in an inundating difference between experimental groups. The day 2 GAG release was highly related with day 1, evaluated with linear regression (R 2 = .935, Figure S2).
The NO, IL-6, and IL-8 release data did not show such inter-donor variation ( Figure S2). Therefore, the results of GAG release from the inflammatory model experiments ( Figure 4D) were analyzed with normalized relative fold changes instead of using the original absolute content.
TNF-α induced a nonrecoverable catabolic shift of NP cells even when it was removed from the medium at 24 hours after supplementation, which is consistent with previous studies. 38 pain and worst back pain. 66 However, Etanercept injection in patients with chronic LBP, more than 6 months' duration, was unable to resolve chronic discogenic pain. 67 Hence, anti-inflammatory treatment with Etanercept at early onset of disc inflammation may be beneficial to relieve discogenic pain by reversing the degenerative cascade. There seems to be a time-point dependent window of therapeutic applicability for anti-inflammation strategies. However, radicular pain indicates IVD herniation, which is a different entity from chronic LBP related to IVD degeneration and may therefore intrinsically respond differently to anti-inflammatory treatment. In clinics, patients are usually treated at a certain period after an acute inflammation or during chronic inflammation process. At this stage, targeting or removal of the inflammatory factor may not be sufficient. Also, treatment to prevent continuous degeneration needs to be included as well.
Limitations: This study solely focused on TNF-α induced acute inflammation within IVDs. Other proinflammatory factors such as IL-1β and lipopolysaccharide may also be used for the same purpose, while the differences in the effects of various factors need to be further evaluated. Both in vitro and ex vivo experiments were only performed within 1 week. Therefore, further studies should be designed to investigate the effect of prolonged or repeated stimulation of TNF-α. The exogenous dose of TNF-α in the current study is much higher than in vivo pathological conditions, and a high dose of TNF-α can induce cell apoptosis and senescenc, which play important roles in IVD degeneration.. 32,38,68 In rat NP cells cultured with TNF-α at 50 ng/mL for 12 hours apoptosis was induced. 68 Also in IVD organs cultured with 200 ng/mL TNF-α for 21 days cell senescence was induced. 38 Further study is warranted to evaluate the effect of TNF-α on cell apoptosis and senescence in long-term within the current model in the future.

| CONCLUSION
The present work sought to address several specific questions on the