The Expression of Toll-like Receptors in Cartilage Endplate Cells: A Role of Toll-like Receptor 2 in Pro-Inflammatory and Pro-Catabolic Gene Expression

Abstract

Introduction: The vertebral cartilage endplate (CEP), crucial for intervertebral disc health, is prone to degeneration linked to chronic low back pain, disc degeneration, and Modic changes (MC). While it is known that disc cells express toll-like receptors (TLRs) that recognize pathogen- and damage-associated molecular patterns (PAMPs and DAMPs), it is unclear if CEP cells (CEPCs) share this trait. The CEP has a higher cell density than the disc, making CEPCs an important contributor. This study aimed to identify TLRs on CEPCs and their role in pro-inflammatory and catabolic gene expression.

Methods: Gene expression of TLR1-10 was measured in human CEPs and expanded CEPCs using quantitative polymerase chain reaction. Additionally, surface TLR expression was measured in CEPs grouped into non-MC and MC. CEPCs were stimulated with tumor necrosis factor alpha, interleukin 1 beta, small-molecule TLR agonists, or the 30 kDa N-terminal fibronectin fragment. TLR2 signaling was inhibited with TL2-C29, and TLR2 protein expression was measured with flow cytometry.

Results: Ex vivo analysis found all 10 TLRs expressed, while cultured CEPCs lost TLR8 and TLR9 expression. TLR2 expression was significantly increased in MC1 CEPCs, and its expression increased significantly after pro-inflammatory stimulation. Stimulation of the TLR2/6 heterodimer upregulated TLR2 protein expression. The TLR2/1 and TLR2/6 ligands upregulated pro-inflammatory genes and matrix metalloproteases (MMP1, MMP3, and MMP13), and TLR2 inhibition inhibited their upregulation. Endplate resorptive capacity of TLR2 activation was confirmed in a CEP explant model.

Conclusions: The expression of TLR1-10 in CEPCs suggests that the CEP is susceptible to PAMP and DAMP stimulation. Enhanced TLR2 expression in MC1, and generally in CEPCs under inflammatory conditions, has pro-inflammatory and pro-catabolic effects, suggesting a potential role in disc degeneration and MC.

Keywords: Modic changes; cartilage endplate; cartilage endplate cells; disc degeneration; toll-like receptors.

Figure 1

Ex vivo analysis of TLR gene expression in CEP tissue. (A–J) TLR1–10 were measured with qPCR in RNA isolated from freshly isolated CEP tissue, HEK cells, and THP1 cells. The graphs present the 2^-DC_T values normalized to GAPDH. Significant differences in expression levels between CEP tissue, THP1 cells, and HEK cells were tested using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test. * p < 0.05, ** p < 0.01, *** p < 0.001. (K,L) Immunohistochemistry stains for (K) TLR2 and (L) TLR4 from the CEP of three representative patients as well as the isotype control within a single patient example are shown (right column). Positive staining was found mainly on the side of the CEP closer to the nucleus pulposus. Positive staining is indicated by brown coloration and nuclei counterstain with Hematoxylin. Scale bars = 50 µm as indicated in the central image; zoomed-in regions are indicated with rectangles and positive stainings are highlighted by arrows.

Figure 2

Expression of surface TLRs involved in DAMP and PAMP recognition; specifically, (A) TLR1, (B) TLR2, (C) TLR4, and (D) TLR6 were measured in unstimulated cultured CEPCs isolated from CEPs adjacent to a degenerated disc with adjacent MC1 (n = 6) and from CEPs adjacent to a degenerated disc without MC1 (non-MC) (n = 9). The graphs illustrate mean fold change ± standard deviation. Significance was tested using unpaired t-tests on log₂ fold change values normalized to the mean of non-MC DCt values. The graphs indicate mean fold change values ± standard deviation. The bold p-values indicate significance (p < 0.05).

Figure 3

In vitro CEPC TLR expression and regulation. (A–D) Gene expression of TLRs in CEPCs following 24 h and 48 h stimulation with varying concentrations of Pam2csk4, Pam3csk4 (10 ng/mL and 1000 ng/mL), LPS (10 ng/mL and 50 ng/mL), FN fragment 30 kDA (FNf 30kDa) (2.5 µg/mL and 5 µg/mL), TNF-α (10 ng/mL), and IL-1β (10 ng/mL) is depicted. The panels illustrate the expression levels of (A) TLR1, (B) TLR2, (C) TLR4, and (D) TLR6. Significance was tested on log₂ fold change in DDC_T values normalized to unstimulated CEPCs by repeated measures one-way ANOVA, followed by multiple comparisons which compared each condition to the unstimulated condition at the respective timepoint. The graphs indicate mean fold change values ± standard deviation. Dunnett’s statistical hypothesis testing was applied to correct for multiple comparisons. The asterisks signify significance: * p < 0.05, ** p < 0.001, *** p < 0.0001.

Figure 4

Inflammatory gene expression of (A) IL-6, (B) IL-8, and (C) CCL2 in CEPCs following 24 h and 48 h stimulation with varying concentrations of Pam2csk4, Pam3csk4 (10 ng/mL and 1000 ng/mL), LPS (10 ng/mL and 50 ng/mL), FN fragment 30 kDa (FNf30 kDa) (2.5 µg/mL and 5 µg/mL), TNF-α (10 ng/mL), and IL-1β (10 ng/mL) is depicted. The graphs indicate mean fold change values ± standard deviation. Significance was tested on log₂ fold change in D D CT values by repeated measures one-way ANOVA, followed by multiple comparisons which compared each condition to the unstimulated condition at the respective timepoint. Dunnett’s statistical hypothesis testing was applied to correct for multiple comparisons. The asterisks signify significance: * p < 0.05, ** p < 0.001, *** p < 0.0001.

Figure 5

Protease gene expression of (A) MMP1, (B) MMP3, (C) MMP9, and (D) MMP13 in CEPCs after either 24 or 48 h of Pam2csk4, Pam3csk4 (10 ng/mL and 1000 ng/mL), LPS (10 ng/mL and 50 ng/mL), FN fragment 30 kDA (FNf) (2.5 µg/mL and 5 µg/mL), TNF-α (10 ng/mL), and IL-1β (10 ng/mL) stimulation. The graphs indicate mean fold change values ± standard deviation. Significance was tested on log2 fold change in D D CT values normalized to unstimulated CEPCs by repeated measures one-way ANOVA, followed by multiple comparisons which compared each condition to the unstimulated condition at the respective timepoint. Dunnett’s statistical hypothesis testing was applied to correct for multiple comparisons. The asterisks signify significance: * p < 0.05, ** p < 0.001, *** p < 0.0001.

Figure 6

Protein expression of TLR2. (A) A representative case of TLR2 measured on CEPCs (blue), with HEK cells (red) serving as the negative control and THP1 cells (orange) serving as the positive control. (B,C) The effect of 72 h of stimulation with (B) Pam2csk4 and (C) Pam3csk4 at concentrations of 10 ng/mL and 1000 ng/mL on TLR2 levels illustrated as percentages normalized to unstimulated CEPCs. Significance was tested with one-way ANOVA on median fluorescence corrected for multiple comparisons using Dunnett’s multiple comparisons test. ** p < 0.01.

Figure 7

Inhibition of TLR2 signaling with three different dosages (50 µM, 100 µM, and 200 µM) of TL2-C29 added 2 h prior to adding Pam2csk4. Genes that showed upregulation through Pam2csk4 stimulation were used to test if this upregulation could be inhibited by blocking TLR2. The inflammatory gene (A) IL-6, the responding TLR (B) TLR2, as well as the proteases (C) MMP1, (D) MMP3 and (E) MMP13 were measured. The graphs indicate mean fold change values ± standard deviation. Statistical significance was tested on log₂ fold change in DDC_T values by repeated measures one-way ANOVA, followed by multiple comparisons which compared each condition to the unstimulated condition at the respective timepoint. Dunnett’s statistical hypothesis testing was applied to correct for multiple comparisons. The asterisks signify significance: * p < 0.05, ** p < 0.001, *** p < 0.0001.

Figure 8

CEP explant model with half the CEP punch biopsy exposed to Pam2csk4. (A) The percentage of released glycosaminoglycans (GAGs) is illustrated, with each pair representing one endplate punch biopsy cut in half. Significance was tested using the Wilcoxon paired one-tailed t-test. (B) The transcriptome of CEP after 14 days of incubation with 10 µg/mL Pam2csk4 is illustrated as log₂ fold change relative to the unstimulated half. Significance was tested using one-way ANOVA, corrected with the two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli. * p < 0.05, *** p < 0.001.