This is the first systematic study showing that CEPC express all TLR1-10. Interestingly, TLR8 and TLR9 expression is lost when the cells are cultured and expanded. Expression of the TLR2 and TLR7 genes have previously been reported using micro arrays [46]. Single cell RNA sequencing and bulk RNA sequencing showed that CEPC are transcriptionally different from nucleus pulposus and annulus fibrosus cells of the intervertebral disc, and hence CEPC should be investigated separately [46], [47]. Disc cells express TLR1-6, 9, and 10 [25]. Therefore, expression of TLR7 seem to be specific to CEPC. In contrast to disc cells, expression of TLRs is independent of the degree of disc degeneration, indicating that regulation of TLR expression is different than in the disc. This underscores that CEPC should be investigated separately from disc cells.
CEPC expressed all cell surface TLRs, i.e. TLR1, TLR2, TLR4, TLR5, and TLR6. Of those, TLR2 and TLR4 are of particular importance in DD and MC, because they can be activated by extracellular matrix fragments generated during DD and by C. acnes, a bacterium that has been found at increased concentrations in discs at MC levels, and that can trigger DD and MC in animal model [26], [36], [37], [40], [48], [49]. TLR2 signaling requires heterodimerization with TLR1 or TLR6 [26]. TLR4 homodimerizes for activating downstream signaling. To test the regulation of TLR1, 2, 4, and 6 under inflammatory conditions, CEPC were stimulated with the TLR1/2 and TLR2/6 specific ligands Pam3csk4 and Pam2csk4, with TLR4 specific ultra-pure LPS, and with IL-1β or TNF-α to test TLR-signaling independent regulation [50]. TLR2 expression was found to be upregulated under all these conditions. TLR4 and TLR6 were not affected at all, and TLR1 was only upregulated by IL-1β. On protein level, stimulation with Pam2csk4 but not with Pam3csk4 enhanced TLR2 expression, indicating a positive feedback loop of TLR2/6 stimulation with TLR2 expression. The correlation of TLR2 expression with TLR6 but not with TLR1 further supports the relevance of TLR2/6 signaling in CEPC. In intervertebral disc cells and chondrocytes from osteoarthritic joints, similar to our data, only TLR2 was upregulated after IL-1β stimulation [50], [51]. Together, this suggests that TLR2/6 signaling could be of importance in CEPC. Whether this affects sensitivity or intensity of TLR2/6 signaling in-vivo cannot be concluded from these experiments.
Expression of pro-inflammatory cytokines and MMPs is enhanced after TLR2 and TLR4 signaling in the disc and in hyaline cartilage, leading to matrix resorption and impaired function [27], [28], [50], [51], [52], [53], [54]. Here, stimulation of TLR1/2, TLR2/6, and TLR4 enhanced expression of pro-inflammatory cytokines and of MMPs, suggesting a potential detrimental role of TLR2 and TLR4 signaling in endplate degradation [17]. Blocking TLR2 signaling abrogated the TLR2/6 enhanced expression of pro-inflammatory cytokines and MMPs. This proves that signaling indeed occurs through TLR2 and that the observed upregulation of cytokines and MMPs is functionally linked to TLR2 stimulation.
TLR2 signaling could occur in CEPC through DAMPs or PAMPs that have been shown to be present in degenerating intervertebral discs. For examples, FNf30 kDa generated during DD activates TLR2 [27], [30], [31], and C. acnes engage TLR2 [38]. Which C. acnes factors activate TLR2 is unclear. In keratinocytes during acne vulgaris, Christie-Atkins-Munch-Petersen 1 factor (CAMP1) binds TLR2 [55], and di- and tri-acylated lipoproteins engage TLR1/2 and TLR2/6 heterodimers, respectively [56]. C. acnes also secretes hyaluronidase, which can cleave hyaluronic acid into fragments acting as DAMPs [28], [57], and sialidases, which can disrupt TLR-inhibitory mechanisms [55], [58]. While it can be expected that CEPC respond to C. acnes, this needs to be proven.
A TLR2 response in the CEP may have other consequences than a TLR2 response in the disc because sensory nerve fibers end in the CEP [59], and hence pro-inflammatory and neurotrophic factors secreted by CEPC after TLR2 stimulation could directly act in a paracrine manner on these fibers and enhance pain sensitization. Importantly, nerve fiber density in CEPs is increased in MC and therefore TLR2 signaling in MC CEPs could affect pain sensitization even stronger [60], [61]. Overexpression of MMPs could also degrade the matrix of the anyway thin CEP and increase the risk for damaging with detrimental consequences like MC [62], avulsion-type herniations [63], Schmorl’s nodes [64], and endplate fractures [65]. Lastly, overexpression of pro-inflammatory cytokines and MMPs by CEPC may also affect the adjacent bone marrow and lead to inflammatory changes even before the CEP is damaged.
This study is limited by the fact that the significance of TLR signalling in CEPC for the pathogenesis of DD and MC cannot be derived from the experiments performed