Hedgehog proteins and parathyroid hormone‐related protein are involved in intervertebral disc maturation, degeneration, and calcification

Abstract Parathyroid hormone‐related protein (PTHrP) and hedgehog signaling play an important role in chondrocyte development, (hypertrophic) differentiation, and/or calcification, but their role in intervertebral disc (IVD) degeneration is unknown. Better understanding their involvement may provide therapeutic clues for low back pain due to IVD degeneration. Therefore, this study aimed to explore the role of PTHrP and hedgehog proteins in postnatal canine and human IVDs during the aging/degenerative process. The expression of PTHrP, hedgehog proteins and related receptors was studied during the natural loss of the notochordal cell (NC) phenotype during IVD maturation using tissue samples and de‐differentiation in vitro and degeneration by real‐time quantitative polymerase chain reaction (RT‐qPCR) and immunohistochemistry. Correlations between their expression and calcification levels (Alizarin Red S staining) were determined. In addition, the effect of PTHrP and hedgehog proteins on canine and human chondrocyte‐like cells (CLCs) was determined in vitro focusing on the propensity to induce calcification. The expression of PTHrP, its receptor (PTHR1) and hedgehog receptors decreased during loss of the NC phenotype. N‐terminal (active) hedgehog (Indian hedgehog/Sonic hedgehog) protein expression did not change during maturation or degeneration, whereas expression of PTHrP, PTHR1 and hedgehog receptors increased during IVD degeneration. Hedgehog and PTHR1 immunopositivity were increased in nucleus pulposus tissue with abundant vs no/low calcification. In vitro, hedgehog proteins facilitated calcification in CLCs, whereas PTHrP did not affect calcification levels. In conclusion, hedgehog and PTHrP expression is present in healthy and degenerated IVDs. Hedgehog proteins had the propensity to induce calcification in CLCs from degenerated IVDs, indicating that in the future, inhibiting hedgehog signaling could be an approach to inhibit calcification during IVD degeneration.


| INTRODUCTION
Over 80% of the human population experiences low back pain during their lifetime. 1 An important cause for low back pain is degeneration of the intervertebral disc (IVD). 2 The IVD consists of an inner gelatinous nucleus pulposus (NP) and outer fibrous annulus fibrosus. During IVD maturation, large, vacuolated notochordal cells (NCs) are replaced by smaller, non-vacuolated chondrocyte-like cells (CLCs), a process that in the human NP has come to near completion already at birth. 3 Thus, a "healthy" human NP from a child or young adult contains CLCs, and often no NCs, implying that the NC transition towards a CLC cellular phenotype is initially part of a maturation process. The healthy NP contains a high glycosaminoglycan (GAG) content that attracts water, whereas during IVD degeneration, the NP GAG and water content decreases and denatured collagen content increases, 4 resulting in NP dehydration and more rigid extracellular matrix (ECM). The avascular IVD exhibits inadequate repair, leading to a vicious circle: the IVD weakens and is more vulnerable to damage by physiologic loading. Consequently, loss of mechanical function, traumatic damage, and pain develop. 5 Current treatments for IVD disease are primarily aimed at relieving symptoms. Therefore, there is an urgent need for agents stimulating biological IVD repair. 6 To develop such treatments, further knowledge of the pathogenesis of IVD degeneration is required.
Numerous signaling pathways have been proposed to play a role in IVD degeneration, including Wnt/β-catenin, Sonic hedgehog (SHH), and hypoxia-inducible factor signaling. [7][8][9] SHH and Indian hedgehog (IHH) morphogens belong to the hedgehog family, which plays a crucial role in embryonic development. Hedgehog proteins undergo intramolecular cleavage catalyzed by their C-terminal domain, yielding an N-terminal product that represents the mature and biologically active hedgehog form and a C-terminal product with no known signalingrelated function. 10 SHH has been related to IVD formation and maintenance 11 and based on this precedence, it is thought to be primarily involved in (patho)physiology of the IVD. In contrast, IHH is considered to primarily have a biologic role at the growth plate and joint cartilage level. 12 While the regulation of hedgehog signaling is well described at the postnatal growth plate and to some extent at the joint cartilage level, the role of these two hedgehogs in postnatal IVDs of species that can suffer from clinical IVD disease remains largely elusive.
At the growth plate level, IHH and PTHrP form a growthrestraining feedback loop regulating chondrocyte differentiation during endochondral ossification. [13][14][15] IHH, produced by prehypertrophic growth plate chondrocytes, promotes proliferation and differentiation, and stimulates calcification independently from PTHrP. 16 13,14,[16][17][18] PTHrP, produced by periarticular chondrocytes, prevents proliferative cells from leaving the proliferating growth plate zone. Moreover, targeted overexpression of PTHrP delays the appearance of hypertrophic chondrocytes. 19 When chondrocytes are no longer sufficiently stimulated by PTHrP, they stop proliferating and synthesize IHH. 20 Hedgehogs have been studied to some extent in joint cartilage degeneration. 21 Considering the high homology of the biologically active N-terminus of IHH and SHH (ie, 88% at the amino acid level) and the fact that these ligands share downstream signaling pathways, the reported studies are not always able to discern which of the hedgehog ligands are involved. A positive correlation between IHH or SHH expression and osteoarthritis (OA) has been described. [22][23][24][25] Furthermore, activated hedgehog signaling aggravated the OA phenotype in genetically modified mice. 23 In line with this, disrupted (Indian) hedgehog signaling prevented hypertrophic chondrocyte differentiation and osteophyte formation in OA cartilage. 26,27 Furthermore, SHH may be involved in redifferentiation of transplanted chondrocytes, as de-differentiated chondrocytes transfected with SHH showed improved cartilage repair in a rat articular cartilage defect model. 28 Notably, PTHR1 expression was decreased in OA chondrocytes 29 and PTHrP suppressed chondrocyte mineralization and hypertrophy, 30 indicating that PTH(rP) signaling might prevent the development and progression of OA. 31 Previous work indicated that IVD degeneration resembles OA, for example, hypertrophic differentiation and calcification are phenomena that can occur in both tissues during degeneration. 32 Although hedgehog and PTHrP are extensively studied in (OA) joint cartilage, limited information is available regarding their role in the postnatal IVD. Previous work indicated that hedgehog 9,33,34 and PTCH1 9,35 are expressed in the postnatal murine IVD, whereas PTHrP was absent. 9 PTH partially recovered NC numbers in the NPs of ovariectomized rat 36 and suppressed calcification in degenerated human CLCs. 37 Thus, results from limited work conducted thus far imply that hedgehog proteins and PTHrP could potentially be targets for IVD repair. To take this a step further, in the present study, the expression of PTHrP and hedgehog proteins was determined in IVD maturation and degeneration by employing postnatal IVDs that can suffer from clinical IVD disease. The setup of the study is presented in Figure 1, and a schematic overview of the results is given in Table 1.

| Study design
In this study, hedgehog, PTHrP, and related receptor expression and function were determined in two different phases: during the loss of the NC phenotype (maturation phase) and in maturated until severely degenerated IVDs (degeneration phase) ( Figure 1). Canine tissues and cells were employed considering the fact that they are readily available, and that dogs experience back pain and IVD degeneration with similar characteristics as humans and are a suitable translational animal model. [38][39][40] Dog breeds can be classified as chondrodystrophic (CD) or non-chondrodystrophic (NCD). 41 CD and NCD dogs demonstrate distinct differences in clinical IVD disease which correlate with their physical appearance and are defined by their genetic background. In CD dogs, NCs are replaced by CLCs around 1 year of age and back pain due to IVD degeneration usually develops around 3 to 7 years of age.
In NCD dogs, NCs can remain the predominant cell type during life and if low back pain develops, it occurs around 6 to 8 years of age. 41 To determine the expression of possible relevant hedgehogs (ie, IHH and SHH) and PTHrP during loss of the NC phenotype, hedgehogand PTHrP-related gene and protein expression was studied in healthymaturated canine IVDs and NCs from healthy NCD canine IVDs. To determine the role of hedgehog and PTHrP in IVD degeneration, F I G U R E 1 Schematic setup of the study. The role of hedgehog proteins (Indian and Sonic hedgehog; IHH and SHH, respectively) and parathyroid hormone-related protein (PTHrP) in the intervertebral disc (IVD) was determined in two different phases: during loss of the notochordal (NC) phenotype (maturation phase) and in maturated until severely degenerated IVDs (degeneration phase). Hedgehog-and PTHrPrelated protein expression was determined using immunohistochemistry in canine and human NPs from healthy until severely degenerated IVDs, for example, natural IVD maturation and degeneration. To further elucidate the expression of hedgehog proteins and PTHrP during loss of the NC phenotype, IHH-and PTHrP-related gene expression was studied in NCs from healthy canine IVDs that lost their specific phenotype and marker (brachyury, cytokeratin 8) expression during monolayer culture. To determine the role of hedgehog proteins and PTHrP in IVD degeneration, the correlation between PTHrP and hedgehog (receptor) expression and calcification was furthermore determined in surgically removed human NP samples across the range of histological IVD degeneration. Lastly, the effect of hedgehog proteins and PTHrP was determined on calcification of canine and human CLCs in vitro T A B L E 1 Hedgehog, PTHrP, and related receptor expression during IVD maturation, degeneration and calcification Note: Hedgehog, PTHrP, and related receptor expression was studied during intervertebral disc (IVD) maturation (loss of the notochordal (NC) phenotype) using different models as given in Figure 1. Hedgehog signaling and PTHrP-related protein expression was determined using immunohistochemistry (IHC) in nucleus pulposus (NP) tissue from healthy (Thompson score I) until maturated (Thompson score II) canine IVDs, for example, natural IVD maturation. Additionally, hedgehog-and PTHrP-related mRNA expression was studied in NCs from healthy canine IVDs that de-differentiated and lost their typical vacuolated phenotype and characteristics during monolayer culture. Furthermore, hedgehog, PTHrP and related receptor expression was studied during IVD degeneration using immunohistochemistry in NPs from maturated (Thompson score II) until severely degenerated (Thompson score V) canine and human IVDs collected during standard postmortem diagnostics. Lastly, the correlation between IVD calcification levels and hedgehog, PTHrP and related receptor expression was studied in surgically removed human NP tissue.  was performed and analyzed as described previously 3 (Supporting information 1). For determination of relative quantitative gene expression, the Normfirst (E ΔΔCq ) method was used. For each target gene, the C qvalue of the test sample was normalized to the mean C q -value of four stably expressed reference genes (GAPDH, HPRT, RPS19, and SDHA):   (Positively stained) cell numbers in each NP were manually counted.
Adobe Photoshop CS6 was used to manually count (positively stained) cell numbers in four (canine) or six (human) randomly selected NP areas per IVD section as described previously. 50 The mean percentage of cells that stained positive over the total number of cells present (ratio) was determined per Thompson score for every target protein.

| Expression of hedgehog proteins and PTHrP during IVD maturation
To study the IVD maturation phase (defined by loss of the NC phenotype), hedgehog and PTHrP signaling-related expression was studied in vitro in NCD canine NC-rich NP tissue and 0 to 4 day cultured NC clusters described to de-differentiate and lose their specific phenotype and marker (brachyury, cytokeratin 8) expression by gene expression analysis 42 and by immunohistochemistry of the NP.
F I G U R E 2 Hedgehog and PTHrP signaling-related mRNA expression in canine nucleus pulposus (NP) tissue and cultured notochordal cell (NC) clusters. The NC clusters de-differentiate, that is, lose their typical vacuolated phenotype and characteristics (eg, brachyury and cytokeratin 8 mRNA expression) during 4-day monolayer culture. mRNA expression of both the hedgehog and PTHrP signaling pathway decreases with NC de-differentiation. IHH, Indian hedgehog; NCD, non-chondrodystrophic; n.s., not significant; PTCH1, patched, PTHrP, parathyroid hormonerelated protein, PTHR1, PTHrP receptor; SHH, Sonic hedgehog; SMO, smoothened. n = 6 (NP tissue) -8 (NC clusters). T0, T2, T4: NC clusters at culture day 0, 2, and 4, respectively. *P < .05 and **P < .01 3.1.1 | Gene expression of canine NC cluster culture IHH messenger RNA (mRNA) expression was significantly higher in NCD canine NP tissue than in 0-to 4-day cultured NCs (P < 0.05), where it was hardly detectable (Figure 2A). SHH mRNA expression was significantly higher in non-cultured NCs (culture day 0) and decreased rapidly during culture (P < .05, Figure 2B). PTCH1 mRNA expression was significantly lower in NCD NP tissue than in 0-to4-day cultured NCs (P < .01, Figure 2C). SMO mRNA expression decreased from day 0 until day 2 and thereafter increased again, and was significantly higher in 4-day cultured NCs than in NCD NP tissue (P < .05, Figure 2G). mRNA of transcription factor GLI1, target gene of hedgehog signaling, was hardly detectable at T0 and T2, and was not differentially expressed between groups ( Figure 2D). GLI2 mRNA was expressed at significantly higher levels in NCD NP tissue than in 0-to 4-day cultured NCs (P < .01), where it was hardly detectable ( Figure 2E). mRNA expression of GLI3 was significantly lower in NCD NP tissue than in 0-to -4 -day cultured NCs (P < .05, Figure 2F). Taken together, although PTCH1 and SMO mRNA was expressed, mRNA of IHH, SHH (both hedgehog ligands) and their transcription factors GLI1 and GLI2 was hardly detectable (C q -values > 35) in canine NCs that lost their typical vacuolated morphology during culture.
PTHrP mRNA expression was significantly higher in freshly isolated NCs than in native tissue (P < .05) and decreased rapidly during NC culture (P < .05, Figure 2H). Also, PTHR1 mRNA expression decreased from day 0 to days 2 and 4 in culture (P < .01, Figure 2I).
Furthermore, it was significantly lower in 2-and 4-day cultured NCs than in NCD NP tissue (P < .05).

| Histology of canine and human IVD maturation and degeneration
In healthy, Thompson score I canine NPs, only NC clusters were detected ( Figure 3E). In maturated, Thompson score II canine NPs, mainly CLCs were present; only one NCD canine NP also contained some NCs. In Thompson scores III to V canine NPs, only CLCs were encountered in single cells and clusters. In NPs from healthy, Thompson score I human IVDs, mainly CLCs and only some NCs were detected, mainly as single cells, but also in clusters ( Figure 4G). In NPs from Thompson scores II to V human IVDs, only CLCs were encountered in single cells and clusters. In both species, CLC clusters were more abundant and larger in NPs from Thompson score IV and V IVDs. Immunopositivity, regardless of the proteins and species, was identified in both single cells and clusters without an evident spatial distribution pattern in both species.

| Hedgehog signaling and PTHrP immunopositivity in canine IVD maturation
The N-terminal hedgehog antibody demonstrating active hedgehog (IHH, SHH) signaling did not render specific staining in the canine samples and therefore, this antibody was only studied in human tissues. Immunopositivity (ratio of positive cells) for PTCH1 and SMO was significantly higher in canine NP tissues from healthy, Thompson score I NC-rich IVDs than in NP tissue from maturated, Thompson score II CLC-rich IVDs (P < .05, Figure 3A,B), suggesting higher expression in NCs than CLCs. PTHrP and PTHR1 immunopositivity did not differ between NPs from healthy and maturated canine IVDs ( Figure 3C,D). Notably, PTHrP immunopositivity in healthy canine IVDs seemed to depend on age: PTHrP was detected in more than 10% of the NCs of 7-and 16-month-old canine donors, whereas this was 0% to 5% in older (17-until 96-month-old) canine donors ( Figure 3C,E) Taken together, the mRNA expression pattern of both hedgehog and PTHrP-signaling related genes in de-differentiated NCs ( Figure 2) were confirmed by the protein expression pattern of maturing canine IVDs (Figure 3), that is, reduced mRNA levels and immunopositivity coincided with the loss of the NC phenotype.

| Hedgehog and PTHrP in canine and human degenerating IVDs
Regardless the species, hedgehog and PTHrP staining was cytoplasmic, while for PTCH1, SMO, and PTHR1, both the cell membrane and the cytoplasm demonstrated immunopositivity (Figures 3E and 4G).
The latter seems to be counterintuitive considering the fact that PTCH1, SMO, and PTHR1 are transmembrane proteins. Nonetheless, cytoplasmic immunopositivity can be well explained for all these three proteins. Hedgehog ligands bind to their transmembrane receptor PTCH1 and cause PTCH1 inactivation, internationalization, and further degradation. Inhibition of PTCH1 is followed by SMO activation (located intracellularly) and translocation to the cell membrane, which initiates the canonical hedgehog signal transduction. 57 As such, dependent on its localization, different forms of PTCH1 can be distinguished: (a) unliganded PTCH1, which is active and localized to the cell membrane and (b) PTCH1 bound to the ligand that is further F I G U R E 3 Hedgehog and PTHrP signaling-related immunopositivity in canine intervertebral discs with different degeneration grades. Both signaling pathways are activated in healthy IVDs and increase with degeneration. The N-terminal hedgehog antibody demonstrating active hedgehog signaling did not render specific staining in the canine samples and was therefore not included. CD, chondrodystrophic; NCD, nonchondrodystrophic; PTCH1, patched; PTHrP, parathyroid hormone-related protein, PTHR1, PTHrP receptor 1; SMO, smoothened. n = 7-8 per Thompson score. *P < .05, **P < .01 and ***P < .001 internalized and may be further degraded. 58 Internalization of PTCH1 is postulated to also contribute to modulation of hedgehog signaling, which remains to be further resolved. The epitope of the PTHR1 antibody maps near the N-terminus which has been demonstrated to remain associated with its ligand within endosomes, challenging the classical paradigm of G protein-coupled receptor signaling. 59 Considering the semi-quantitative nature of immunostainings and technical challenges concerning quantification, follow-up analysis of immunostainings did not distinguish between the different patterns of immunopositivity.

| Degenerating canine IVDs
Considering the differential clinical representation of IVD disease among dog breeds, 41 the canine IVDs were separately analyzed for CD and NCD dogs. PTCH1 immunopositivity, a direct target of active hedgehog signaling, was significantly higher in NPs from Thompson scores IV to V canine IVDs than in NPs from Thompson score II to III IVDs (P < .05, Figure 3A). Only NCD dogs showed a positive correlation between PTCH1 immunopositivity and macroscopic IVD degeneration grade (Thompson scores II-V, P < .05, r:0.552), whereas only CD dogs showed a positive correlation between SMO immunopositivity and macroscopic IVD degeneration grade (P < .05, r:0.621). PTHrP immunopositivity significantly increased during canine IVD degeneration (P < .01, Figure 3C). In both CD and NCD dogs, a positive correlation between PTHrP immunopositivity and macroscopic IVD degeneration grade was observed (P < .05, r:0.627 [CD] and P < .01, r:0.748 [NCD]). Immunopositivity for PTHR1 increased from NPs of maturated, Thompson scores II to III IVDs to severely degenerated, Thompson score V IVDs (P < .05, Figure 3D).
Only CD dogs showed a positive correlation between PTHR1 immunopositivity and macroscopic IVD degeneration grade (P < .01, r:0.696). Although the current study demonstrates some differences in hedgehog-PTHrP (related) expression between CD and NCD dogs, which may be related with their differential genetic background, the direction of the correlation between protein expression and IVD degeneration grade was similar.

| Human degenerating IVDs
As expected, a strong, positive correlation was found between the macroscopic Thompson and the histologic IVD degeneration score 48,49 of the human samples collected postmortem (r:0.907, P < .001; Figure 4F). To be able to relate these results to the follow-

| Hedgehog and PTHrP in IVD calcification
To further explore the role of hedgehog and PTHrP in more clinically relevant material, immunopositivity was correlated to calcification in surgically removed human NP samples across the range of histological IVD degeneration. 48,49 Lastly, the effect of IHH and SHH was deter-

| Calcium levels increase in the human NP during IVD degeneration
Based on their well-described role in chondrocytes, 22 30 In contrast, in the present study, PTHrP immunopositivity was similar between NPs with low or abundant calcification and PTHrP did not affect calcification of CLCs in vitro. However, assuming that PTHrP inhibits the IVD calcification process in vivo, the absence of PTHrP upregulation in later stages of degeneration even in the presence of high PTHR1 levels may allow for NP calcification in the process of degeneration.

| SHH and IHH induce calcification in canine and human CLCs
In contrast to PTHrP, IHH is known to promote chondrocyte hypertrophy and calcification in OA. [22][23][24] To the best of our knowledge, this is the first study to demonstrate that IHH and especially SHH also have the propensity to induce calcification in human and canine CLCs from degenerated IVDs in vitro. Interestingly, SHH was even more potent in inducing calcification in human and canine CLCs than IHH.
Together with the increased hedgehog expression in IVDs with high levels of calcium, this may suggest a possible role for SHH and/or IHH in calcification during IVD degeneration. In line with this thought, disrupted IHH signaling prevented hypertrophic chondrocyte differentiation and osteophyte formation in OA cartilage. 26,27 Possibly, this could also be a therapeutic approach to prevent or retard these processes during IVD degeneration. From a clinical perspective, however, additional considerations need to be acknowledged, including possible side effects of inhibited hedgehog signaling, for example, on the cardiovascular 66 and central nervous 67 system.