Intervertebral disc degeneration is one of the most prominent risk factors for chronic low back and neck pain, the leading cause of global disability, and a major contributor to the opioid addiction crisis (Global Burden of Disease Study, 2013 Collaborators, 2015; Deyo et al., 2015; GBD, 2016 Disease and Injury Incidence and Prevalence Collaborators, 2017). The nucleus pulposus (NP) cells reside at the center of the avascular, and thus hypoxic niche of the intervertebral disc (Rajpurohit et al., 2002; Risbud et al., 2006) and have adapted to this hypoxic environment by controlling the expression and activity of hypoxia-inducible factors (HIFs) (Fujita et al., 2012a; Fujita et al., 2012b). Two isoforms of HIF-α most prominently expressed in the intervertebral disc are HIF-1α, and HIF-2α that bind to the identical hypoxia response element (HRE) motif and dimerize with a common binding partner HIF-1β/ARNT (Loboda et al., 2010). However, prior studies have demonstrated that there are cell type and tissue type-specific differences between HIF-1α and HIF-2α function and their transcriptomic targets (Hu et al., 2003; Downes et al., 2018). A broader categorization of their functions shows that HIF-1α primarily controls cell metabolic activities, whereas HIF-2α regulates cell proliferation, survival, extracellular matrix homeostasis, and oxidative defenses (Downes et al., 2018).

While both HIF-α isoforms are robustly expressed in the NP, HIF-1α is the critical driver of glycolysis, energetics, and overall survival of cells within the specialized niche of the disc (Agrawal et al., 2007; Silagi et al., 2018; Madhu et al., 2020; Johnston et al., 2022). Notochord-specific FoxA2Cre; HIF-1αf/f mice show a prominent NP cell apoptosis by birth and subsequently, the NP tissue is replaced by cartilage-like tissue, highlighting the central role of HIF-1α in survival and adaptation of NP cells (Merceron et al., 2014). Noteworthy, in contrast to the dramatic phenotype of HIF-1αFoxA2Cre mice and the pivotal role of HIF-1 in NP cells, our recent studies of HIF-2αFoxA2Cre delineating the role of HIF-2α in the developing disc pointed to a different function of this HIF homologue. HIF-2αFoxA2Cre mice did not result in early cell death, but showed mild and transient protection against age-dependent intervertebral disc degeneration characterized by reduced NP tissue fibrosis (Johnston et al., 2022). While these investigations implied a pro-pathogenic role of HIF-2α in the disc, little is known about the consequences of elevated HIF-2α activity in the NP compartment and its impact on overall intervertebral disc health. This is particularly relevant since increased HIF-2α levels and activity have been seen in human degenerated discs (Huang et al., 2019). Therefore, the goals of this study were twofold, to establish a causal link between elevated HIF-2α activity and disc health in the aging mouse spine and to identify potential biological pathways responsive to HIF-2α induction in the NP, in vivo.

Clinically, humans with Epas1 gain-of-function syndrome fail to prune or have normal venous regression leading to vascular malformations (Rosenblum et al., 2021). Notably, mice generated with the homologous Epas1 mutation showed similar vascular malformations (Wang et al., 2019). HIF-2α overexpression in adipocytes is shown to promote cardiac hypertrophy (Lin et al., 2013). Likewise, increased HIF-2 is known to promote cancer cell proliferation (Gordan et al., 2007), and patients with elevated HIF-2 have poor prognosis (Gao et al., 2017). These data suggest the overall pathological consequences of increased HIF-2 activity in different disease contexts. Concerning musculoskeletal pathologies, in addition to markedly elevated levels in degenerated human discs and enhanced expression of catabolic markers following HIF-2α expression in cultured human NP cells (Huang et al., 2019), HIF-2α levels are shown to be higher in osteoarthritic human and mouse cartilage, and Epas1 haploinsufficiency is enough to lessen the severity of traumatic osteoarthritis in mice (Saito et al., 2010; Saito and Kawaguchi, 2010). Corroborating these findings; Yang and others showed that induction of HIF-2α levels in knee cartilage of mice with Ad-Epas1 increased cartilage destruction (Yang et al., 2010). Some of the effects of HIF-2α on arthritis severity are likely mediated by promoting chondrocyte apoptosis and expression of matrix-degrading enzymes (Ryu et al., 2012). Interestingly, in contrast to these findings, in vitro studies of human articular chondrocytes showed the importance of HIF-2α in promoting the expression of healthy cartilage matrix genes (Lafont et al., 2007; Lafont et al., 2008). Likewise, in the Tibetan population, HIF-2α polymorphisms that decrease HIF-2 transcriptional activity and are thus critical in adaptation to a low-oxygen environment, a higher incidence of osteoarthritis and lower back pain has been noted (Hoy et al., 2003; Liang et al., 2018). Together these results imply that the HIF- 2α function in the skeleton is tissue and context-dependent (Saito et al., 2010; Saito and Kawaguchi, 2010; Yang et al., 2010).

Based on our findings of HIF-2αFoxA2Cre mice, elevated levels seen in degenerated human discs, and its known role in osteoarthritis in mice, we hypothesized that increased HIF-2α expression in the NP would compromise disc health as the spine ages. Our results show that elevated HIF-2α activity causes degenerative changes in the NP compartment prominently evident in the aged spine. Our transcriptomic studies showed that elevated HIF-2α activity resulted in the upregulation of genes enriched in biological themes related to primary cilia, cell motility, and axonal guidance by SLIT/ROBO, whereas, genes in themes related to GLP-1, aldolase, and TCA cycle were downregulated reflecting an overall down-modulation of cell metabolism. Supporting our earlier loss-of-function studies in HIF-2αFoxA2Cre mice and based on the findings of this investigation, we conclude that elevated levels of HIF-2α in the NP are one of the pathogenic mechanisms contributing to intervertebral disc degeneration with spine aging.

All mouse experiments were performed under protocols approved by the Institutional Animal Care and Use Committee (IACUC) of Thomas Jefferson University following the relevant guidelines and regulations. Krt19tm1(cre/ERT)Ggu/J (Keratin19CreERT or K19CreERT) mice were purchased from Jackson Laboratories. It has been shown K19CreERT allele shows a robust recombination in the NP compartment following tamoxifen injections from P8 up to 15 months of age (Mohanty et al., 2020). The Gt(ROSA)26Sortm4(HIF2A*)Kael/J (HIF-2αdPA, on C57BL6/J background) developed by Dr. William Kaelin were used in the study. HIF-2αdPA (dPA) mice have hemagglutinin-tagged human HIF2A cDNA modified with two proline to alanine substitutions (P405A, P531A) introduced 3′of the mouse Gt(ROSA)26Sor promoter in the HIF-2α/Epas1 gene (Kim et al., 2006) thus promoting the stability of HIF-2α. Male and female control (dPA), homozygous (K19CreERT; HIF-2αdPA or K19-dPA) and heterozygous (K19CreERT; HIF-2αdPA/+ or K19-dPA/+) littermate mice were generated and injected with Tamoxifen once a day for three consecutive days at 3 months of age as described before and the mice were analyzed at 14- and 24-month of age (Johnston et al., 2022).

Spines dissected en bloc were fixed in 4% PFA for 48 h at 4 °C and decalcified in 20% EDTA at 4 °C for 21 days, washed with 1× PBS, and then placed in 70% ethanol before paraffin embedding. Lumbar discs were sectioned in the coronal plane at 7 μm and then stained with 1% Safranin O, 0.05% Fast Green, and 1% hematoxylin. The sections were visualized with Axio Imager 2 microscope (Carl Zeiss) using a 5×/0.15 N-Achroplan or 20 ×/0.5 EC Plan-Neofluar (Carl Zeiss) objective. Images were captured with Axiocam 105 color camera (Carl Zeiss) using Zen2 software (Carl Zeiss). Histological grading of 14 months (n = 11 dPA control, 9 K19-dPA/+, 7 K19-dPA) mice, four to five lumbar discs/mouse, 33–48 lumbar discs/genotype) and 24-month-old mice (n = 9 dPA control, 10 K19-dPA/+, 7 K19-dPA mice, four to five lumbar discs/mouse, 33–51 lumbar discs/genotype) was performed by three blinded graders using modified Thompson grading scale and Tessier scale for endplate scoring (Tessier et al., 2020a; Tessier et al., 2020b; Gorth et al., 2020; Novais et al., 2021). Since the unique interactions between biological, biomechanical, and genetic factors at individual spinal levels have been shown to produce different phenotypic outcomes, each disc was considered as an independent sample (Tessier et al., 2020a; Tessier et al., 2020b; Tsingas et al., 2020; Novais et al., 2021).

μCT imaging was performed on lumbar spines of 14- and 24-month-old mice (n = 6 dPA and 6 K19-dPA; 6 lumbar discs and 7 vertebrae/mouse were analyzed) using the high-resolution μCT scanner (SkyScan 1,275, Bruker, Konitch, Belgium). Samples were placed in 1× PBS and scanned with an energy of 50 kV and current of 200 μA, resulting in 15 μm3 voxel size resolution. Intervertebral disc height and the length of the vertebral bones were measured and averaged along the dorsal, midline, and ventral regions in the sagittal plane. Disc height index (DHI) was calculated as previously described (Tsingas et al., 2020; Novais et al., 2021; Johnston et al., 2022).

Picrosirius Red staining and analysis were performed on 7 μm mid-coronal disc sections of 14- month-old mice (n = 11 dPA control, 9 K19-dPA/+, 7 K19-dPA, four to five lumbar discs/mouse, 33–48 lumbar discs/genotype) and 24-months-old mice (n = 9 dPA, 10 K19-dPA/+, 7 K19-dPA, four to five lumbar discs/mouse, 33–51 lumbar discs/genotype) as described before (Choi et al., 2018). The stained sections were imaged using Eclipse LV100 POL (Nikon, Tokyo, Japan) with a 10×/0.25 Pol/WD 7.0 objective and Nikon’s Digital Sight DS-Fi2 camera. Under polarized light, stained collagen bundles appear as either green, yellow, or red pixels that correlate to fiber thickness indicating thin, intermediate, or thick fiber thickness, respectively (Dayan et al., 1989). Using NIS Elements Viewer software (Nikon, Tokyo, Japan), fibers were quantified by thresholding for green, yellow, and red pixels over the selected region of interest (ROI) for NP and AF.

Mid-coronal disc sections (7 μm) were deparaffinized in histoclear and rehydrated in graded ethanol solutions (100%–70%), water, and 1× PBS. Antigen retrieval was performed using either citrate-buffer, proteinase-K, or MOM kit (Vector Laboratories, Burlingame, CA, USA; BMK-2202) depending on the antibody. Sections were blocked for 1 h, at room temperature, in 5%–10% normal goat or donkey serum as appropriate in PBS-T (0.4% Triton X-100 in PBS) or using the reagent from the MOM kit. Then, they were incubated at 4 °C overnight in primary antibodies against anti-HA (1:100; BioLegend; 901,533), CA3 (1:150; Santa Cruz Biotechnology; sc-50715), CS (1:300; Abcam; ab11570), ACAN (1:50; Millipore; ab1031), ARGxx (1:200; Abcam; ab3773), COMP (1:200, Abcam; ab231977), COL I (1:100; Millipore; abt123), COL II (1:400; Fitzgerald; 70R-CR008), After incubation with primary antibody, tissues were washed and reacted with Alexa Fluro-594 (Ex: 591 nm, Em: 614 nm) conjugated secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) at a 1:700 dilution in blocking buffer for 1 h at room temperature. The sections were then washed and mounted with ProLong Gold Antifade Mountant with DAPI (Thermo Fisher Scientific, P36934). Slides were visualized with the Axio Imager 2 (Carl Zeiss Microscopy) using 5×/0.15 N-Achroplan (Carl Zeiss Microscopy) or 10×/0.3 EC Plan-Neofluar (Carl Zeiss Microscopy) or 20×/0.5 EC Plan-Neofluar objectives, X-Cite 120Q Excitation Light Source (Excelitas Technologies), AxioCam MRm camera (Carl Zeiss Microscopy), and Zen2 software (Carl Zeiss Microscopy). All quantifications were conducted in 8-bit grayscale using the Fiji package of ImageJ (Schindelin et al., 2012). Images were thresholded to create binary images, and NP and AF compartments were manually segmented using the Freehand Tool. These defined regions of interest were analyzed either using the Analyze Particles (cell number quantification) function or the Area Fraction measurement.

FTIR spectral acquisition and analysis utilizing deparaffinized sections (7 μm) of decalcified mouse lumbar disc tissues from control and mutant animals at the 14- and 24-month (n = 6 mice/genotype/age, 2 disc/mouse, 12 total discs/genotype) were performed as described previously (Tsingas et al., 2020; Novais et al., 2021). Spectrum Spotlight 400 FT-IR Imaging system was used to acquire spatial resolution images in the mid-IR region from 4,000 to 800 cm−1 (Perkin Elmer, Waltham, MA, USA) of three consecutive sections/disc to minimize section-based variation. Isys Chemical Imaging Analysis software (v. 5.0.0.14) was used to analyze the mean second-derivative absorbances which were quantified and compared between the genotypes. Additionally, collected spectra were analyzed using K-means clustering analysis in the Eigenvector Solo + MIA software (v. 8.8) to agnostically delineate anatomical regions within the disc. Regions of IR images are separated into two or more classes, or “clusters,” according to spectral similarity. The K-means in our analyses was K = 7. During each iteration, the remaining objects (pixels of the spectral image) are assigned to one of these clusters based on the distance from each of the K targets. New cluster targets are then calculated as the means of the objects in each cluster, and the procedure is repeated until no objects are reassigned after the updated mean calculations.

Statistical analysis was performed using Prism 9 (GraphPad, La Jolla, CA, USA) with data presented as whisker box plots showing all data points with median and interquartile range and maximum and minimum values. Differences between distributions were checked for normality using Shapiro-Wilk tests and further analyzed using an unpaired t-test (two-tailed) for normally distributed data and the Mann-Whitney U test for non-normally distributed data. Comparison of multiple groups of normally distributed data was performed using one-way ANOVA, whereas non-normally distributed data were analyzed using Kruskal–Wallis and Dunn’s or Welch’s multiple comparison tests. Analyses of Modified Thompson Grading distributions and fiber thickness distributions were performed using a chi-square test at a 0.05 level of significance. Since the interactions between genetic, biological, and biomechanical factors at individual spinal levels have been shown to produce different phenotypic outcomes in humans and mice, each disc was considered an independent sample (Battié et al., 2008; Li et al., 2016; Vincent et al., 2019; Tessier et al., 2020a; Tsingas et al., 2020).

We have previously shown that the deletion of HIF-2α in the NP compartment (HIF-2αFoxA2Cre) causes transient protection from early degenerative changes in 14-month-old, middle-aged mice (Johnston et al., 2022). We therefore hypothesized that elevated HIF-2α levels and activity in NP will accelerate age-dependent degenerative changes. To test this hypothesis, we conditionally expressed HIF-2α (HIF-2αdPA) in NP that is resistant to prolyl hydroxylase-mediate degradation and studied the disc phenotype of mice as a function of spine aging. Accordingly, HIF-2αdPA mice were crossed with an NP-targeting tamoxifen-inducible K19CreERT driver, and the resulting littermate dPA control, K19-dPA/+, and K19-dPA mice were injected for three consecutive days at 3-month (3M) and aged up to 24-month (24M) (Figure 1A). To confirm the expression of HA-tagged HIF-2αdPA protein in the disc, we stained the disc sections with an anti-HA antibody which showed cellular staining in the NP compartment of K19-dPA but not in the control dPA mice (Figure 1B).

FIGURE 1. Elevated HIF-2α levels in the NP cause age-dependent disc degeneration in mice (A) Timeline of experimental mice generation and analysis. (B) K19-dPA mice show expression of HA-tagged HIF-2dPA in the lumbar discs of 14-month (14M) dPA and K19-dPA mice stained with anti-HA antibody, n = 2 discs/animal, 3 animals/genotype (scale bar = 100 μm; insert scale bar = 25 μm). (C) Safranin O/Fast Green staining of lumbar discs showing disc morphology and overall proteoglycan content in the intervertebral disc in 14M control (dPA), heterozygous (K19-dPA/+), and homozygous (K19-dPA) animals (scale bar = 200 μm). (D) Histological grading assessment of 14M lumbar using the modified Thompson scale. (E) Safranin O staining of lumbar discs in 24M animals. Yellow arrowheads indicate the loss of cell band and fibrotic remodeling in the NP along with the loss of demarcation between NP and AF compartments. Scale bars, top row: = 200 μm, bottom rows: 50 μm (F) Histological grading assessment of NP and AF using the modified Thompson scale. ns = not significant (G) Comparison of histological grading scores as a function of aging within each genotype. (H) Endplate grading assessment at 14M and 24M. n = 14M: 11 dPA (4 males, 7 females), 9 K19-dPA/+ (3 males, 6 females), 7 K19-dPA (5 males, 2 females) mice, four to five lumbar discs/mouse, 33–48 lumbar discs/genotype and 24M: n = 9 dPA (4 males, 5 females), 10 K19-dPA/+ (5 males, 5 females), 7 K19-dPA mice (4 males, 3 females), four to five lumbar discs/mouse, 33–51 lumbar discs/genotype. The significance for grading distribution was determined using a chi-square test. The significance of differences among 3 or more groups was determined using one-way ANOVA or Kruskal–Wallis with Dunn’s test. Quantitative measurements represent the median with the interquartile range.

To understand if increased HIF-2α levels impact intervertebral disc morphology, sections were stained with Safranin O, Fast Green, and Hematoxylin. Histological assessment of 14-month-old (14M) mice showed no apparent differences in NP and AF compartment morphology, and overall disc structural integrity was comparable between the genotypes with no signs of disc herniation (Figure 1C). At this age, a majority of discs, regardless of the mouse genotype showed vacuolated NP cells surrounded by proteoglycan-rich extracellular matrix. The endplate appeared normal, and the AF showed concentric lamellae interspersed with fibrocartilaginous AF cells. However, as expected, NP and AF compartments of some discs showed signs of early degeneration with smaller NP cell band size and distortion of AF lamellae along with disruption of the NP-AF tissue interface. However, these aging-associated changes were present across all the genotypes (Supplementary Figure S1A). Accordingly, at 14 months, the NP and AF average grades between genotypes were comparable, and the grade distributions showed only small differences with mutant mice showing a slight reduction in the percentage of healthy grade 1 discs (Figure 1D). At 24 months (24M), however, there was a small but significant increase in NP average grades of degeneration in K19-dPA mice (Figure 1E). This was also captured by the changes in the grade distributions showing a pronounced increase in the percentage of highly degenerated grade 5 discs with severe annular defects and contained herniations (Figure 1E). When the grading data within each genotype was compared across the ages, the increase in average grades of degeneration was more pronounced in K19-dPA mice compared to dPA control animals in both the NP and AF compartments (Figures 1F,G). Scoring of the cartilaginous endplates (EP) using the Tessier grading scale (Tessier et al., 2020a) showed no significant alterations in EP scores between the individual genotypes across the age (14 M vs. 24 M) and between the genotypes at each of the time points (Figure 1H, Supplementary Figure S1B). Although, the combined average grades of all genotypes at 24 M are slightly higher than at 14M likely reflecting a small aging effect.

To assess if there were corresponding changes in disc hydration-related parameters, microCT measurements of disc height (DH), and vertebral body height (VBH), were performed. At both 14M and 24M, K19-dPA mice showed no differences in DH or VBH. However, when disc height measurements were normalized with vertebral lengths to calculate the disc height index (DHI), at 14M, there was a small but significant decrease in DHI compared to control mice which was attenuated as the mice aged to 24M (Figures 2A,B, Supplementary Figure S2). These analyses showed that elevated expression of HIF-2α in NP cells may promote mild degenerative changes in the disc and affect disc hydration.

FIGURE 2. Increased HIF-2α activity in NP results in a decreased disc height index. (A) Representative reconstructed μCT images of hemi-sections of lumbar motion segments in 14- and 24-month-old dPA and K19-dPA mice, scale bar = 1 mm. Lumbar disc height, vertebral height, and disc height index (DHI) in (B) 14M and 24M dPA and K19-dPA mice. n = 14M: 6 dPA (3 males, 3 females), 6 K19-dPA (4 males, 2 females) mice; 24M: 6 dPA (3 males, 3 females), 6 K19-dPA (3 males, 3 females) mice; 6 lumbar discs and 7 vertebrae/mouse were analyzed. Significance for quantitative measures was determined by using Mann–Whitney U test or an unpaired t-test with Welch’s correction, as appropriate. Quantitative measurements represent the median with the interquartile range.

To understand if the overexpression of HIF-2dPA in the NP affected collagen turnover in the AF compartment, the organization of the collagen matrix was assessed using Picrosirius Red staining coupled with polarized light microscopy. Polarized imaging revealed that at 14M, there were fewer thin and more intermediate fibers in the AF indicating less collagen turnover in the AF compartment of K19-dPA mice (Figures 3A–C). However, at 24M, the differences noted between the genotypes in both quantity and distribution at 14M were not apparent. These analyses suggest that increased HIF-2α in the NP affects AF collagen turnover during early aging but that these changes are secondary to the effects of spine aging.

FIGURE 3. K19-dPA mice evidence altered collagen fibrillar composition and compositional changes in NP and AF. (A–C) Representative polarized images of Picrosirius Red-stained lumbar disc sections (A) and quantification of collagen fiber thicknesses (B, C) performed on 14M and 24M dPA and K19-dPA mice (scale bar in A = 200 μm). 14M: n = 11 dPA (4 males, 7 females) 7 K19-dPA mice (5 males, 2 females), four to five lumbar discs/mouse, 33–48 lumbar discs/genotype and 24M: n = 9 dPA (4 males, 5 females), 7 K19-dPA (4 males, 3 females) mice, four to five lumbar discs/mouse, 33–51 lumbar discs/genotype. (D) Spectral cluster analysis images of 14 M and 24 M discs (scale bar = 200 μm). (E) Average superimposed second derivative spectra, inverted for positive visualization of the NP and AF of 14M and 24M dPA and K19-dPA mice. 14M (F) and 24M (G) quantification of mean second derivative peaks associated with chondroitin sulfate (1,064 cm−1), polysaccharide-associated peak (1,080 cm-1), amide III (1,230 cm-1), collagen (1,338 cm−1), and polysaccharide associated peak (1,032 cm-1). AU = arbitrary units. For both 14M and 24M, n = 6 mice/genotype, 3 discs/mouse, 18 total discs/genotype. Significance for quantitative measures was determined by using Mann–Whitney U test or an unpaired t-test with Welch’s correction, as appropriate. Quantitative measurements represent the median with the interquartile range.

Since there were changes in AF collagen fiber content, Imaging-Fourier transform infrared (FTIR) spectroscopy was used to assess matrix compositional changes. K-means clustering was used to define broader anatomical regions of the disc using chemical compositions. The K-means clustering showed similarly defined regions between dPA and K19-dPA at 14M. However, moderate changes in the NP compartment composition were noted at 24M, suggesting broader changes in the intervertebral disc chemical composition (Figure 3D). Interestingly, while NP and AF spectra showed a clear and expected distinction in both genotypes, in K19-dPA mice, more prominently at 24M, AF absorbance peaks between 1,400–1,600 cm-1 showed some congruence with NP absorbance peaks within the same range, indicating diminished compositional distinction between the tissue compartments (Supplementary Figure S3A). When the average spectra were compared between the genotypes, apparent differences in NP absorbance peaks were noticeable in K19-dPA at both 14M and 24M, suggesting there were compositional differences in the NP (Figure 3E). Differences in the absorbance peaks of the AF between genotypes were less noticeable at both ages than NP (Figure 3E). These results indicated that there were molecular changes in overall NP composition in K19-dPA mice.

To evaluate the specific chemical compositional differences in the NP, and AF, peaks at 1,064 cm−1, 1,120 cm−1, 1,228 cm−1 (chondroitin sulfate), 1,338 cm−1 (collagens), 1,660 cm−1 (total protein, Amide I), 1,550 cm−1 (Amide II), 1,230 cm−1 (Amide III), 1,032 cm−1 and 1,080 cm−1 (polysaccharides), and 1,156 cm−1 (cell-associated proteoglycans) were assessed. At 14M in K19-dPA mice, there was a higher amount of collagen and proteoglycan-associated peaks in the AF (Figure 3D). In addition, we assessed a CS-associated peak at 1,064 cm−1, which showed an increase in the AF (Figure 3F). There was also an increase in the polysaccharide-associated peak at 1,080 cm−1 and amide III-associated peak at 1,230 cm−1 in the AF (Figure 3F). Whereas, at 24M, there were decreases in the collagen-associated peak and an increase in the CS-associated peak 1,064 cm−1 in the AF (Figure 3G). There was also an increase in polysaccharide-associated peak at 1,032 cm−1 (Figure 3G). However, there were no differences in amide II-, chondroitin sulfate-associated peaks at 1,120 cm−1 and 1,128 cm−1, proteoglycan-, polysaccharide associated peak 1,080 cm−1, and total protein-associated peaks (Supplementary Figures S3B, C). Together the FTIR data indicate that elevated HIF-2α activity in the NP alters the molecular composition of the disc, which may result in compromised disc matrix functionality and precipitate overall disc degeneration.

To understand if the degenerative phenotype in the K19-dPA was due to changes in NP cell phenotype and major ECM constituents of the disc, we studied their expression pattern and abundance. The abundance of carbonic anhydrase 3 (CA3), an NP phenotypic marker, was comparable between the genotypes at either time point (Figures 4A,B). Similarly, another NP phenotypic marker, glucose transporter 1 (GLUT1), showed no discernible differences between the genotypes at either time points suggesting that the NP cells did not alter their phenotype in K19-dPA mice compared to dPA control animals (Supplementary Figures S4A, B). Interestingly, increased ACAN levels were observed in the NP of 14M K19-dPA relative to dPA; however, this difference was not evident at 24M. Interestingly, despite increased ACAN, CS levels were lower in the NP compartment of 14M K19-dPA mice, an observation consistent with the FTIR findings and implying decreased ACAN substitution. Similarly, consistent with the increased AF proteoglycan and chondroitin sulfate (CS) content seen in FTIR analysis at the 24M, ACAN and CS levels in AF showed increased abundance in K19-dPA at 24M and levels increased as K19-dPA mice aged from 14 to 24 months suggesting transition into more mucinous matrix. On the other hand, ARGxx, which measures ADAMTS-dependent aggrecan degradation, showed no changes between genotypes at both time points. Additionally, K19-dPA mice did not show an age-dependent decrease in the abundance of cartilage oligomeric matrix protein (COMP) noted in the NP compartment of dPA control mice (Figures 4C,D). COL I and COL II in AF also showed negligible changes with a slight increase noted in COLI abundance in the NP of 14M K19-dPA mice. These analyses suggested that while the NP notochordal phenotype is maintained, there were alterations in a few major ECM components of the disc following increased HIF-2α activity.

FIGURE 4. HIF-2α overexpression alters the abundance of select key ECM molecules of the disc. (A–B) Quantitative immunofluorescent staining of 14M and 24M dPA and K19-dPA lumbar discs for carbonic anhydrase 3 (CA3), aggrecan (ACAN), aggrecan neoepitope (ARGxx), and chondroitin sulfate (CS), scale bar CA3, CS = 200 μm, scale bar ACAN, ARGxx = 100 μm. (C–D) Quantitative immunofluorescent staining of 14M and 24M dPA and K19-dPA lumbar discs for cartilage oligomeric matrix protein (COMP), collagen I (COL1), and collagen II (COL2), scale bar = 100 μm, n = 7 mice/genotype/time point, one to three discs/mouse, 7–21 discs/genotype/marker. White dotted lines demarcate disc compartments. Quantitative measurements represent the median with the interquartile range. Significance was determined using Kruskal–Wallis with Dunn’s test.

To evaluate the global transcriptomic changes in the NP, microarray analysis was used to analyze NP tissue RNA from 14- and 24M K19-dPA and control dPA mice. Principal component analysis (PCA) showed that the 14M dPA and K19-dPA clustered separately (Figure 5A). Hierarchical clustering of the DEGs (p ≤ 0.05, fold change (FC) ≥±1.7) was performed; this data is represented as heatmaps and volcano plots (Figures 5B,C). To better interpret these changes, the CompBio tool was used to understand the biological context of up and downregulated DEGs, which showed super clusters comprised of top themes relating to cilia, cell motility, IGF activity regulation, hypoxia, ROBO/SLIT, and neurotrophin TRKB-receptor activity in the upregulated 14M dataset (Figures 5D,E). The enriched DEGs in these upregulated themes included, Dnah11, Hk2, Igfbp5, Nov and Ntrk2 (Figure 5E, Supplementary Figure S5). One super cluster presented in the downregulated dataset contains themes related to lipid metabolism and GLP-1 regulation with decreased expression in genes such as Glp1, Adcy3, Fxyd3, Ppargc1, Rps6ka2 (Figures 5D,E).

FIGURE 5. Transcriptomic analysis shows alterations in several key biological pathways in NP tissue of 14M K19-dPA mice. (A) Three-dimensional Principal component analysis (PCA) showing discrete clustering of mice based on their genotypes, n = 4 mice/genotype (B) Heat map and hierarchical clustering of Z-score of significantly differentially expressed genes (DEGs) from 14M K19-dPA vs. dPA (p ≤ 0.05, ≥1.7-fold change). (C) Log-log volcano plot of DEGs in the NP shows statistical significance (p-value) versus magnitude of change (fold change). (D) CompBio analysis of DEGs and associated upregulated and downregulated themes in a ball and stick model. The enrichment of themes is shown by the size of the ball and connectedness is shown based on thickness of the lines between them. Themes of interest are colored, and superclusters comprised of related themes are highlighted. (E) Top thematic DEGs plotted based on CompBio entity enrichment score.

At 24M, PCA plots showed a clear separation between control dPA and K19-dPA samples (Figure 6A). Identified DEGs were also plotted as heatmap and volcano plots (Figures 6B,C) and there were more DEGs compared to seen at 14M (Figure 6C). Within the upregulated DEGs, CompBio analysis showed unique superclusters containing themes related to ubiquitin-like hydrolase activity and SCF-mediated degradation of p27/p21 with increased expression of Ufl1, Cul1, Skp2, Fbxw11, and Fbxl5 (Figures 6D,E, Supplementary Figure S6A). Of interest, a select group of enriched DEGs clustered into a theme named “Oxygen-Dependent Proline Hydroxylation of Hypoxia-inducible Factor Alpha” with increased expression shown in Epas1, Cul2, and Sdha (Figures 6D,E). On the other hand, downregulated DEGs showed a supercluster containing themes related to actin dynamics, endodeoxyribonuclease activity, and pyrophosphatase activity with decreased expression shown in Apex, Fen1, Sun2, and Ankrd2. Interestingly, themes named “TCA Cycle” and “Weak Cry” were also found to be downregulated in 24M and contained genes including Asns, Mdh2, Slc16a3, and Rorα amongst the top genes within these themes, respectively (Figures 6D,E, Supplementary Figure S6B).

FIGURE 6. Transcriptomic analysis of NP tissues shows alterations in important biological pathways in 24M K19-dPA mice. (A) Three-dimensional Principal component analysis (PCA) showing discrete clustering of mice based on their genotypes, n = 4 mice/genotype (B) Heat map and hierarchical clustering of Z-score of significantly differentially expressed genes (DEGs) from 14M K19-dPA vs. dPA (p ≤ 0.05, ≥1.7-fold change). (C) Log-log volcano plot of DEGs in the NP shows statistical significance (p-value) versus magnitude of change (fold change). (D) CompBio analysis of DEGs and associated upregulated and downregulated themes in a ball and stick model. The enrichment of themes is shown by the size of the ball and connectedness is shown based on thickness of the lines between them. Themes of interest are colored, and superclusters comprised of related themes are highlighted. (E) Top thematic up and downregulated DEGs plotted based on CompBio entity enrichment score.

Notably, at 14M and 24M, in the upregulated DEGs there was overlap with the thematic superclusters related to motile cilia, HIF/hypoxia, and SLIT/ROBO signaling. Motile cilia themes contained upregulated genes including Dnah11, Rsph4a, Cfap69, and Pacrg. The HIF/hypoxia themes had increased expression of Egln3, Pgf, Ankrd37, Adm, and Edn2. Whereas, the themes SLIT/ROBO signaling included common entities: Slit1, Raph1, and Elmod1. When 14M and 24M datasets were further analyzed to evaluate highly significant common genes irrespective of their thematic clustering (FDR ≤0.05, FC ≥ ±2.0), 11 genes including Pgf, Ntrk2, Pifo, Sez6l, Fabp5, Podn, Slit1, Nme5, Elmod1, Drc1 and Plppr1 were noted (Supplementary Figure S7). These common DEGs are plausible HIF-2 target genes responsive to elevated HIF-2α activity, however, this hypothesis requires additional functional validation outside the scope of the present investigation. Overall, these studies strongly suggest that increased HIF-2α levels in NP alters several key biological pathways linked to cilium, cell motility, and metabolic functions.

Together our studies clearly provide evidence that elevated HIF-2α levels in the NP are pathogenic and alter many key parameters of disc health in the aging spine (Figure 7).

FIGURE 7. Schematic summarizing the consequences of elevated HIF-2α levels in the NP compartment on intervertebral disc health with spine aging. Elevated HIF-2α expression in the NP compartment promotes disc degeneration in mice with aging. The discs of K19-dPA mice show alterations in ECM composition and collagen turnover, Transcriptomic analysis of differentially expressed genes showed their enrichment in biological themes/concepts that are related to cilia, SLIT/ROBO pathway, and HIF/Hypoxia signaling along with unique themes not shared between 14- and 24 months of age. Created with BioRender.com.

This study provides clear evidence that elevated HIF-2α activity in the NP causes age-dependent, mild degenerative changes in the disc corroborating our previous investigation that showed transient protection from age-related disc degeneration in mice with NP-specific deletion of HIF-2α (Johnston et al., 2022). Importantly, these results underscore a possible causal link between elevated HIF-2α levels and disc degeneration seen in humans (Huang et al., 2019). It is important to note that while HIF-2α is expressed in the NP throughout the post-natal life and appears to be dispensable for maintaining the overall health of NP cells but rather pathogenic in the context of spine aging (Johnston et al., 2022) It may still play other regulatory functions under specific circumstances. For example, as is noted in the growth plate and articular chondrocytes, HIF-2 may contribute to NP cell differentiation/hypertrophy or cell death in the context of disc injury, which we have not addressed in the present study (Stewart et al., 2006; Ryu et al., 2012; Wang et al., 2019).

The histological scoring of the NP compartment of 24-month-old K19-dPA mice showed a small but significant increase in scores compared to their littermate control mice. While the grading scores did not differ significantly, a significant decrease in disc height index (DHI) was noted at 14 months suggesting that the degenerative cascade and changes in the water-binding ability of the disc matrix were initiated earlier. This conclusion was supported by the alterations noted in the overall composition of the K19-dPA intervertebral discs. Supporting the moderate nature of phenotypic changes in the disc resulting from changes in HIF-2α levels, similar changes in the tissue morphology and matrix composition have been noted in the limb bud mesenchyme as well as the NP following the loss of HIF-2α (Araldi et al., 2011; Johnston et al., 2022).

The transcriptomic analysis of the NP tissue provided further insights into the pathways that were affected by the elevated HIF-2α activity in the NP and underscored the degenerative phenotype. The findings indicated that HIF-2α in the NP may upregulate the glycolytic enzyme Hk2 at 14 months, which hints at a plausible role in HIF-2α affecting metabolic aspects, a finding that was implicated in a recent loss-of-function HIF-2α study (Johnston et al., 2022). Although HIF-1α is known to play the central role in regulating glycolytic metabolism, HIF-2α has been shown to have a unique metabolic role with increased activity of this transcription factor causing downregulation of lipid metabolism and glucagon regulation at 14 months. Notably, Glp1, glucagon-like peptide 1, is also shown to have strong anti-inflammatory functions in many tissues (Nauck, 2016). Importantly, GLP-1R is expressed in cartilage (Andersen et al., 2021) and its agonist liraglutide is shown to exert beneficial effects on joint tissue health in the context of osteoarthritis (Meurot et al., 2022). Therefore, downregulation of Glp1 likely indicates perturbations in tissue-level anti-inflammatory mechanisms in K19-dPA mice and may in part underscore progressive degenerative changes.

The transcriptomic analysis at 24M provided further insights that metabolic alterations may underlie the observed phenotypic changes. Of interest, there was a supercluster containing one upregulated theme related to “oxygen-dependent hydroxylation of HIF-α”. This theme contained increased expression of Epas1, which encodes for HIF-2α, Egln3, and Sdha, succinate dehydrogenase complex A, suggesting increased mitochondrial oxidation and ROS generation in 24M NP of K19-dPA, likely contributing to the mild degenerative phenotype. Noteworthy, our lab previously showed that HIF-1α-deficient cells decreased TCA cycle flux to succinate, and lessened SDH activity, in an effort to increase flux to glutamate to maintain redox balance in NP cells (Madhu et al., 2020). Whereas increased, Egln3, or PHD3, could implicate impaired oxygen sensing mechanisms, as this protein serves as co-activator of HIF-1 and loss-of-function of Egln3 in mice has been shown to promote disc degeneration (Schoepflin et al., 2017). In terms of downregulated superclusters, there was one theme contained in a cluster related to the TCA cycle with top entities including Asna, asparagine synthetase, which can modulate mitochondrial response (Liu and Birsoy, 2021), and decreased Mdh2, malate dehydrogenase 2, suggesting less formation of oxaloacetate and slowed TCA cycle activity. This theme also highlighted the downregulation of Slc16a3, or MCT4, a plasma-membrane monocarboxylate which is essential to maintain a balance between glycolytic and TCA cycle flux as its inhibition leads to disc degeneration (Silagi et al., 2020). In this context, perhaps HIF-2α may indirectly regulate NP cell metabolism and transport mechanisms, as previously implicated for glucose and sodium-dependent transport in our HIF-2α loss-of-function HIF-2αFoxA2Cre mouse model (Johnston et al., 2022).

Interestingly, there was another downregulated theme that contained the entity Rorα, RAR-related orphan receptor alpha, known to be a major circadian clock gene. We have shown that there is an interaction between BMAL1 and RORα and they modulate HIF-1α activity as well as ECM homeostasis and perturbation in circadian clock genes results in disc degeneration (Suyama et al., 2016; Dudek et al., 2017; Wang et al., 2022; Dudek et al., 2023). Notably, corroborating our current findings, HIF-2αFoxA2Cre mouse has shown upregulation in circadian clock pathways. These results suggested that perturbations of cell metabolism and circadian clock may partially underscore matrix changes and degeneration in K19-HIF-2dPA mice. Of interest, there was a recent study highlighting contribution of elevated HIF-2α levels to increased contractility and pathological transformation of pericytes into vascular smooth muscle-like cells in pulmonary arterial hypertension, and its activation aids in vascular remodeling (Kim et al., 2024).

Although the 14M and 24M transcriptomic data showed some unique differences, there were commonalities in their transcriptional signatures including themes related to cilia, cell motility, hypoxia, and SLIT/ROBO signaling, which were all upregulated. Our lab has shown that primary cilia alter their length in response to changes in extracellular osmolarity in the NP, thus suggesting a relationship between water-binding matrix molecules and ciliary function (Choi et al., 2019). Moreover, in response to mechanical loading, cilia are shown to maintain PTH-PTHR1 signaling axis in the NP cells to promote TGFβ1 signaling and aggrecan levels (Zheng et al., 2018). Relevant to our findings in K19-dPA mice, HIF-2α in murine neuronal cells is shown to increase ciliary length in hypoxia by interacting with intraflagellar transport protein 88 homolog (IFT88) (Leu et al., 2023). Upregulation in the SLIT/ROBO cluster presented increased expression of Slit1 which regulates secreted glycoproteins that bind to ROBO receptors known to mediate cell migration (Tong et al., 2019). An upregulation of SLIT-ROBO signaling was also observed within the NP after BNIP3 knockdown which showed dysregulation of metabolic function (Madhu et al., 2023).

Identification of highly enriched common DEGs at 14- and 24-month, uncovered additional biological insights. The placental growth factor (Pgf) was the most upregulated common DEG, this gene is known to be expressed by ischemic or damaged tissues and has pro-angiogenic functions (Rajakumar et al., 2001; Newell and Holtan, 2017). It was therefore not unreasonable to assume that NP cells experienced stress from HIF-2α overexpression and Pgf was a compensatory response to mitigate this stress. Similarly, neurotrophic receptor tyrosine kinase 2, Ntrk2, was recently proposed as a prognostic gene for intervertebral disc degeneration, supporting its upregulation in this degenerative context (Li et al., 2022). The identification of the HIF-2α responsive functions and genes in AF and the endplate would provide complementary information to understand the compartment-specific functionality of HIF-2α in the disc (Wiesener et al., 2003; Agrawal et al., 2008; Shved et al., 2017). While our studies draw clear distinctions between the non-redundant functions of two HIF-α isoforms in the disc, inhibition of HIF-2α activity in the context of spine aging is likely to provide beneficial effects in maintaining disc health.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm.nih.gov/geo/, GSE249908.

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcell.2024.1360376/full#supplementary-material

Agrawal, A., Gajghate, S., Smith, H., Anderson, D. G., Albert, T. J., Shapiro, I. M., et al. (2008). Cited2 modulates hypoxia-inducible factor-dependent expression of vascular endothelial growth factor in nucleus pulposus cells of the rat intervertebral disc. Arthritis Rheum. 58 (12), 3798–3808. doi:10.1002/art.24073

PubMed Abstract | CrossRef Full Text | Google Scholar

Agrawal, A., Guttapalli, A., Narayan, S., Albert, T. J., Shapiro, I. M., and Risbud, M. V. (2007). Normoxic stabilization of HIF-1alpha drives glycolytic metabolism and regulates aggrecan gene expression in nucleus pulposus cells of the rat intervertebral disk. Am. J. Physiol. Cell Physiol. 293 (2), C621–C631. doi:10.1152/ajpcell.00538.2006

PubMed Abstract | CrossRef Full Text | Google Scholar

Andersen, D. B., Grunddal, K. V., Pedersen, J., Kuhre, R. E., Lund, M. L., Holst, J. J., et al. (2021). Using a reporter mouse to map known and novel sites of GLP-1 receptor expression in peripheral tissues of male mice. Endocrinology 162 (3), 162. doi:10.1210/endocr/bqaa246

CrossRef Full Text | Google Scholar

Araldi, E., Khatri, R., Giaccia, A. J., Simon, M. C., and Schipani, E. (2011). Lack of HIF-2α in limb bud mesenchyme causes a modest and transient delay of endochondral bone development. Nat. Med. 17 (1), 25–26. doi:10.1038/nm0111-25

PubMed Abstract | CrossRef Full Text | Google Scholar

Battié, M. C., Videman, T., Levälahti, E., Gill, K., and Kaprio, J. (2008). Genetic and environmental effects on disc degeneration by phenotype and spinal level: a multivariate twin study. Spine 33 (25), 2801–2808. doi:10.1097/BRS.0b013e31818043b7

PubMed Abstract | CrossRef Full Text | Google Scholar

Choi, H., Madhu, V., Shapiro, I. M., and Risbud, M. V. (2019). Nucleus pulposus primary cilia alter their length in response to changes in extracellular osmolarity but do not control TonEBP-mediated osmoregulation. Sci. Rep. 9 (1), 15469. doi:10.1038/s41598-019-51939-7

PubMed Abstract | CrossRef Full Text | Google Scholar

Choi, H., Tessier, S., Silagi, E. S., Kyada, R., Yousefi, F., Pleshko, N., et al. (2018). A novel mouse model of intervertebral disc degeneration shows altered cell fate and matrix homeostasis. Matrix Biol. 70, 102–122. doi:10.1016/j.matbio.2018.03.019

PubMed Abstract | CrossRef Full Text | Google Scholar

Dayan, D., Hiss, Y., Hirshberg, A., Bubis, J. J., and Wolman, M. (1989). Are the polarization colors of picrosirius red-stained collagen determined only by the diameter of the fibers? Histochemistry 93 (1), 27–29. doi:10.1007/BF00266843

PubMed Abstract | CrossRef Full Text | Google Scholar

Downes, N. L., Laham-Karam, N., Kaikkonen, M. U., and Ylä-Herttuala, S. (2018). Differential but complementary hif1α and hif2α transcriptional regulation. Mol. Ther. 26 (7), 1735–1745. doi:10.1016/j.ymthe.2018.05.004

PubMed Abstract | CrossRef Full Text | Google Scholar

Dudek, M., Morris, H., Rogers, N., Pathiranage, D. R., Raj, S. S., Chan, D., et al. (2023). The clock transcription factor BMAL1 is a key regulator of extracellular matrix homeostasis and cell fate in the intervertebral disc. Matrix Biol. 122, 1–9. doi:10.1016/j.matbio.2023.07.002

PubMed Abstract |