Short‐Term Dynamic Unloading of Bovine Tail Discs in Culture Partially Mitigates Induced Degeneration After One‐Strike Trigger

ABSTRACT

Introduction

Intervertebral disc (IVD) degeneration is driven by a vicious circle of interrelated biological and biomechanical factors. Dynamic unloading, defined as dynamic partial decompression, promotes water and metabolite flow, which is essential for IVD homeostasis. However, the mechanobiological effects of unloading remain poorly understood. IVD organ cultures offer a valuable model for studying IVD degeneration and regeneration at the molecular level. This study investigated the biological and biomechanical effects of induced degeneration and the subsequent short‐term dynamic unloading of bovine tail IVDs in a bioreactor culture system.

Methods

We applied a one‐strike degenerative trigger on Day 0 and assessed its immediate effects after 1 day of culture under bioreactor loading (Timepoint 1). The impact of dynamic unloading for three additional days (Timepoint 2) was evaluated in comparison to continued loading. We evaluated biological outcomes, namely cell viability, gene expression, water/sulfated glycosaminoglycan (sGAG) ratio, and sGAG release. Mechanical readouts included disc height, slope of the elastic zone, area under the curve, and neutral zone characteristics.

Results

On Timepoint 1, we demonstrated degeneration in the nucleus pulposus with altered viability, increased inflammatory and catabolic gene expression, elevated sGAG release, a decreased slope of the elastic zone, and an increased area under the curve. On Timepoint 2, we noticed a sustained degenerative cascade in both degeneration groups. However, unloading showed a trend towards partial mitigation of the induced degeneration with decreased iNOS and TRPV4 expression, an increased water/sGAG ratio, reduced sGAG release, and recovery of the disc height.

Conclusion

This first ex vivo study on unloading mechanobiology of bovine degenerated IVDs unveils encouraging preliminary insights. The findings suggest potential benefits of unloading and, more broadly, therapeutic movement as regenerative strategies for degenerated IVDs. These results underscore the need for further studies and encourage research combining mechanical and biological approaches in organ culture models.

In a bovine intervertebral disc organ culture study including biological and mechanical evaluation parameters, unloading ‐compared to loading‐ partially mitigated the degenerative cascade (Timepoint 2) previously induced with a one‐strike trigger (Timepoint 1).

1. Introduction

Low back pain (LBP), often associated with intervertebral disc (IVD) degeneration, is the leading cause of years lived with disability worldwide, imposing a significant societal and economic burden [1, 2]. Investigating IVD degeneration and exploring regenerative strategies are crucial for addressing LBP in the long‐term.

IVD degeneration is driven by a vicious cycle of biological and biomechanical factors. Although genetic predisposition plays a major role [3, 4] external influences such as smoking or mechanical overload [3, 5, 6], may also contribute to IVD degeneration. In the degenerative state, the cell and extracellular matrix metabolism shifts towards catabolism, characterized by decreased cell viability, increased expression of matrix metalloproteinases (MMP) and collagen I, reduced proteoglycan content, and decreased collagen type II expression [7, 8, 9, 10]. As aggrecan molecules become cleaved and lose their ability to retain water, disc height decreases, leading to altered mechanical properties with reduced intradiscal pressure [11, 12, 13] and diminished mechanical resistance. With mild to moderate degeneration, IVD stiffness decreases, and the neutral zone widens [14, 15, 16]. These mechanical changes exacerbate cellular and extracellular matrix dysregulation, perpetuating the degenerative cycle.

Unloading, defined as dynamic partial decompression of the IVDs, is a natural physiological process that occurs during sleep, when lying down, or adopting specific postures [17]. This process promotes hydration and metabolite flow, which is essential for IVD homeostasis [8]. Increased water content in the IVD has been observed after overnight bedrest, as confirmed by MRI T2 imaging [18]. Similarly, clinical traction protocols have demonstrated an increase in IVD hydration [19], along with reductions in functional disability and pain intensity among patients with degenerated IVDs [19, 20].

Despite these findings, the mechanobiological effects of unloading on the IVD remain poorly understood, particularly in relation to disc phenotype, water and proteoglycan content, and mechanical properties. Studies in animal models (rats and rabbits) have shown promising results, including increased collagen type II expression, proteoglycan content, and disc height [21, 22, 23]. However, extrapolating these results to humans is challenging due to size differences and the presence of persistent notochordal cells, which enhance anabolic turnover of their IVDs compared to human IVDs [24].

IVD organ cultures in bioreactors offer a valuable platform for studying IVD degeneration and regeneration. These models are in line with the 3Rs guidelines [25], allowing precisely controlled IVD loading while enabling biochemical and molecular analyses. Moreover, they facilitate the application of standardized degenerative triggers, such as chemical or mechanical stimuli [26]. Most bioreactor studies on IVD regeneration focus on biomaterials, stem cells, or pharmacological agents [27], aiming to slow down or halt catabolic processes. However, the potential regenerative effects of movement and unloading have received little attention in these models. To address this gap, we previously developed a dedicated bioreactor system with a mechanical testing device specifically designed to study the mechanobiology of active dynamic unloading in bovine tail IVDs [28].

The current study aimed to investigate the biological and biomechanical effects of a one‐strike degenerative trigger and subsequent short‐term dynamic unloading in a bioreactor culture setting. By comparing unloading with continued loading, we examined whether and how mechanical interventions can influence the degenerative cycle in the IVDs. We hypothesized that unloading would positively impact the degenerative IVD by alleviating or halting the degeneration process or even restoring the IVD function.

2. Materials and Methods

2.1. Study Summary

Bovine tail IVDs were exposed to a degenerative trigger by a one‐strike impact loading and subsequently subjected to dynamic unloading. The effects of the degenerative trigger were evaluated on Day 1, whereas the effects of unloading, compared to loading, were assessed on Day 4 through biological and biomechanical outcomes including cell viability, gene expression, water/sGAG ratio, and sGAG release in the medium, as well as disc height, elastic zone slope, area under the curve, and neutral zone characteristics (Figures 1 and 2A).

FIGURE 1.

Study groups represented on a dissected tail illustrating the intervertebral discs (IVDs) harvested (IVD 1–IVD 6). The IVD height (Δx) is the distance between the two growth plates, identified as the widest parts of the IVD segment and indicated, for IVD 4, by the two red arrows.

FIGURE 2.

Study description including the study timeline, the study design, the loading regimes, the modified bioreactor system, with the lid and spring of the intervertebral disc chamber that allows controlled static loading, and the study evaluation parameters. From Day 2 to Day 4, the loading command is represented as starting at 0.00 MPa to facilitate the readability of the illustration; however, due to the technical limitations of the mechanical testing device, we effectively started the loading at −0.01 MPa as described in the text.

2.2. Sample Description and Preparation Including One‐Strike Degenerative Trigger

We used young bovine tail IVDs obtained as byproducts from a local slaughterhouse. No animals were sacrificed specifically for this research. The characteristics of the calves from which the tails were sourced are summarized in Table 1.

TABLE 1.

Characteristics of the bovines whose tails were collected.

Study animal number	Mass (kg)	Age (months)	Sex	Species
1	131	4	Male	Crossed Swiss Brown (Dairy cattle)
2	230	13	Male
3	110	12	Female
4	178	13	Female
5	144	5	Male
6	138	5	Male
7	183	14	Male
8	156	7	Female
9	138	7	Male

We harvested six IVDs per tail and distributed them across six experimental groups (Figure 1): (1) “Day 0”; Timepoint 1 groups (2) “Loading” and (3) “Degeneration”; Timepoint 2 groups (4) “Loading”, (5) “Degeneration + loading” (DegLd) and (6) “Degeneration + unloading” (DegUnld). The largest harvested IVD was assigned to the Day 0 group, as its size often exceeded the bioreactor system's limits. We generally assigned the two smallest harvested IVDs to the Timepoint 1 groups, where the primary goal was to characterize the one‐strike induced degeneration. The mid‐sized IVDs were assigned to Timepoint 2 groups, where the impact of unloading on degenerated IVDs, the main research question of this study, was investigated.

A total of nine different tails (Table 1) were used, providing n = 9/group. To minimize variability, samples were distributed evenly across groups within one Timepoint, ensuring equal representation of IVD sizes (cf. Table S1).

Sample preparation on Day 0 followed previously described protocols [28, 29]. Briefly, tails were washed, disinfected, and dissected before sectioning. Day 0 IVD segments were then cut ¹ through the bony endplates [29]. Culture samples were prepared by cutting through the cranial and caudal vertebral bodies, leaving approximately 10 mm of bone attached to the disc [28]. For IVDs assigned to the degeneration groups (group Degeneration of Timepoint 1, and groups DegLd and DegUnld of Timepoint 2), we applied a one‐strike degenerative trigger, consisting of a 2 min static compressive load of 0.5 MPa followed by a single impact compression of 50% of the disc height (approximately 1 s). To ensure the completion of the one‐strike trigger, we recorded the maximum pressure achieved and evaluated the presence of endplate cracks following impact (Table S1). We labeled the IVD segments on the posterior side of the cranial part of the vertebral body to track segment orientation through the experiment and ensure consistent measurement of IVD area and systematic tissue collection. To facilitate nutrient circulation, we drilled a hole in the vertebral bodies. Samples were then embedded ² and cleaned using a jet lavage system with Ringer solution, washed with a phosphate buffered saline containing penicillin–streptomycin, and stored overnight in the incubator at 37°C, 5% CO₂ and 90% relative humidity in 55 mL of medium under free swelling conditions.

2.3. IVD Culture and Loading Protocol

On Day 1, after overnight free swelling, all the samples of Timepoints 1 and 2 underwent cyclic dynamic compressive loading for 2 h at 0.2 Hz (−0.048 MPa to −0.248 MPa). The starting compression force (−0.048 MPa) was determined based on previous calculations [28] to replicate the compressive force applied by the soft tissues of the spine.

The amplitude of the dynamic compressive loading (0.2 MPa) was determined based on previous studies using physiological loading conditions [29, 30]. We verified the orientation of the IVD segments and mounted all segments in the bioreactor system with the cranial side facing up.

After completing the dynamic loading session on Day 1, we collected the IVD segments of the Timepoint 1 groups (Loading and Degeneration) and the medium of all groups for Timepoint 1 analysis, added fresh culture medium to the Timepoint 2 groups (Loading, DegLd and DegUnld), and applied a basal static compressive loading (−0.048 MPa) to the Timepoint 2 samples to mimic the mechanical constraints imposed by the spine's soft tissues while storing them in the incubator until the next loading session (Figure 2A–C).

In summary, Timepoint 1 evaluated the impact of the one‐strike degenerative trigger, followed by one loading session (Figure 2B).

From Day 2 to Day 4, IVD segments were dynamically loaded for 2 h per day at 0.2 Hz, following the same sequence as on Day 1, with 22 h of static loading between the dynamic loading sessions. The samples from the groups Loading and DegLd were loaded in compression from −0.01 to −0.2 MPa, whereas the samples from the group DegUnld were loaded in traction (or unloading), from −0.01 to 0.024 MPa. These loading parameters were determined based on clinical traction protocols and translated to the bovine IVD model as described previously [28]. Between each session, IVDs remained under static compression (−0.048 MPa) from Day 1 onwards.

On Day 4, after the final dynamic loading or unloading session, we collected all IVD segments and media for Timepoint 2 analysis (Figure 2A–C).

In summary, Timepoint 2 assessed the effect of unloading on degenerated IVDs, comparing it to loading on degenerated and undegenerated discs (Figure 2B).

The loading protocol on Day 1 slightly differed from Days 2 to 4 (Figure 2C). Before Day 1, samples were in free swelling conditions, whereas before Days 2–4, samples were under static compression (−0.048 MPa). To account for this difference, we adjusted the loading protocol and removed the basal compression (−0.048 MPa) from Days 2 to 4. However, due to technical constraints of the mechanical testing device, the loading could not begin at exactly 0 N. Instead, loading for all samples started at −0.01 MPa. To evaluate the accuracy of dynamic loading and unloading, we calculated the RMSE of peak and valley forces in the different groups for Timepoints 1 and 2 from the loading data recorded at 20 Hz, that is 100 points/cycle.

In this study, static loading was applied via a controlled spring compression system (Figure 2C), allowing for variable compressive force. Compared to the previous version of the chamber [28], we redesigned the chamber lid and the static loading lid accessory to be MRI‐compatible for future research.

2.4. Sample Processing

We collected and processed Day 0 samples immediately after harvesting to prevent alterations in water content or gene expression (Figure 2A,B).

The IVD segments assigned to Timepoint 1 were collected directly at the end of the first bioreactor session on Day 1, whereas the Timepoint 2 samples were collected after the final bioreactor session on Day 4 (Figure 2A,B). The sample processing followed previously described protocols [28].

We removed the embedding and the bony endplates. Next, we cut out the frontal right quarter of the IVD and snap froze it in cryocompound for further histological viability and morphological analyses. We removed the cartilage endplates of the leftover disc. We then divided it into the three IVD regions (nucleus pulposus (NP), inner annulus fibrosus (AFi) and outer annulus fibrosus (AFo)) for separate analysis. We respectively froze or snap froze the chopped tissue pieces for further matrix and water content or gene expression analyses. We also froze the medium collected on Day 1 and Day 4.

2.5. Biological Evaluation

We analyzed the sulfated glycosaminoglycan (sGAG) release in the medium by performing a direct spectrophotometric microassay with 1.9 dimethyl‐methylene blue (DMMB) at pH 3 and A530 as previously described [28, 31] (Figure 2D). Additionally, we quantified the nitrite (breakdown product of nitric oxide (NO)) release in the medium with the Griess Reagent System (A530) according to the manufacturer's instructions.

As previously [28], we also determined the water content/sGAG content (mg/mg) ratio in the IVD tissue. The samples were freeze dried for 24 h. The difference in mass between the wet tissue mass on the collection day and the dry tissue mass after freeze drying represented the water content. The proteoglycans were extracted from the dry tissue as reported before [32, 33] and the sGAG in the tissue was quantified via the DMMB assay at pH 1.5 (A525–A595). Moreover, we ran 1.2% agarose gel electrophoresis for selected samples to detect proteoglycan degradation [33, 34]. We loaded samples containing 10 μg sGAG, used bovine fetal epiphyseal isolated aggrecan A1D1 as a positive control, and chondroitin 4‐sulfate sodium salt as a negative control, and stained the gels with toluidine blue.

For the RNA extraction, to improve yield and quality, we combined elements of the previous protocol [35] and used phase maker tubes for the phase separation step. So, we pulverized the frozen samples, homogenized them in TriReagent with carrier, performed the phase separation step with 1‐bromo‐chloro‐propane in phase maker tubes twice, and then continued the procedure with RNeasy spin columns following the manufacturer's instructions. As previously described [28], we executed the reverse transcription and performed the real‐time polymerase chain reaction for RPLP0 (reference gene), COL1A1, COL2A1, ACAN, MMP3, ADAMTS5, iNOS, TRPV4, and TRPC6. We used the gene expression of the reference sample Day 0 from the same tail for normalization. The technical information of the qPCR can be found in Table S3.

We analyzed the viability and the morphology on sagittal IVD cryosections of 5 μm nominal thickness, respectively stained with Lactate Dehydrogenase and Ethidium Homodimer (LDH/Eth) [36] or Weigert's Hematoxylin, Safranin O, and Fast Green [37]. We imaged the slides with an Olympus BX63 upright microscope taking Z‐stacks, which was important for the LDH/Eth sections to capture the entire thickness of the section for LDH and Eth signal. Separately from that, we followed the procedure previously described [28].

2.6. Mechanical Evaluation

We determined the area of the IVD after dissection by measuring the diameters. To improve the accuracy of the measure, we performed the measurements with constant‐force calipers. Additionally, thanks to the oriented labeling of the IVDs, we could systematically and specifically assess the antero‐posterior and latero‐lateral diameters, which improves the reliability of the measures and the subsequent calculation of the loading forces based on the IVD area. To determine the amplitude of the one‐strike, we measured the disc height as the distance between the two growth plates, identified as the parts of the IVD segment with the largest diameter (Figure 1), at the border between the IVD and the vertebral bodies. To follow up on the disc height evolution over time, we measured the segment height in the mid‐sagittal and mid‐frontal planes after dissection, after the one‐strike, and after the free swelling night. Afterward, we measured the disc height before and after the loading and unloading sessions in the bioreactor system, based on the height variation of the stem of the upper part of the IVD holder.

The force‐displacement curves at the beginning (Cycle 30) and at the end (Cycle 1439) of each session in the bioreactor system were evaluated as previously described [28] (Figure 2E). The changes in the slope of the elastic zones, representing stiffness, and in the area under the curve, quantifying hysteresis, were evaluated for all samples over time, not only within a bioreactor session but also between the different bioreactor sessions over the culture days. In the unloading group, as the force range went below and above 0 N, the slope of the neutral zone, as well as the force and displacement range of the neutral zone, was calculated additionally.

2.7. Statistical Analysis

To assess the degenerative impact of the one‐strike trigger, we compared the Loading and Degeneration groups of Timepoint 1. To evaluate the impact of subsequent short‐term dynamic unloading, we compared the three groups of Timepoint 2, namely Loading, Degeneration + loading (DegLd) and Degeneration + unloading (DegUnld). We compared the biological and mechanical evaluation parameters of Timepoints 1 and 2, respectively, with paired t‐tests and one‐way repeated measures ANOVAs with Tukey–Kramer's post hoc tests. We chose a repeated measures design to account for the matched samples, as the IVDs in the groups came from the same tails. On Timepoint 1, the replicates (Figure 2B) of the sGAG and NO measures in the medium and the replicates of the mechanical evaluation parameters were averaged in both groups to allow pairwise comparison of the groups.

Additionally, differences in the overall evolution over culture days of the different mechanical evaluation parameters were tested based on the slope of the linear regression line fitting the various measurement points (Cycles 30 and 1439 of the dynamic bioreactor session over the culture days). Two‐way repeated measures ANOVAs with Tukey–Kramer's post hoc tests were performed to compare the parameters at each measurement point over time in the different groups. A p value < 0.05 was considered significant.

We tested the normal distribution of the data with the D'Agostino‐Pearson, Anderson‐Darling, andShapiro‐Wilk tests and tested the equality of the variances of the differences between the groups via Mauchly's test. Data that did not fit the assumptions for parametric tests were log10 transformed to match the parametric tests' requirements [38]. Nevertheless, in case log10 transformed data did not fit the assumptions either, results of the nonparametric test were reported additionally. In the presence of outliers that were not methodological errors but unusual data points defined by Grubbs' test (α = 0.05), results of the parametric tests were reported with and without outliers. Additionally, we added trend indicators, represented by pink arrows, when a minimum of 6 donors out of 9 followed the same trend between the groups in the absence of a significant p value.

The statistical analysis and its rationale are further elaborated in the Supporting Information.

3. Results

3.1. Reliability of the Loading

The pressure achieved during the one‐strike impact loading was comparable across all groups, with no statistical differences observed (Figure 3). Thus, we considered the baseline degeneration to be consistent across the groups.

FIGURE 3.

Pressure achieved during the 50% disc height compression of the one‐strike on Day 0 was similar in the samples spread over the different groups. The small differences observed between the different groups, indicated by pink arrows, did not lead to differences in the evaluation parameters. The data are represented as mean ± standard deviation (SD).

Given that the degenerative strike was strain‐controlled, variations in compressive pressure were expected across samples. A trend (at least 6 samples out of 9) was observed in the Degeneration group on Timepoint 1, where the mean compressive pressure (−12.5 ± 2.8 MPa) was slightly lower than in the DegLd and DegUnld groups (−14.2 ± 2.0 MPa and −13.7 ± 2.2 MPa). However, this difference did not lead to any significant changes in biological or mechanical outcomes. The trend may be attributed to disc size differences. As a result, we anticipate a stronger degenerative cascade in the DegLd and DegUnld groups on Timepoint 2, ensuring the progression and completion of the degenerative processes in these samples compared to the Degeneration group on Timepoint 1.

We also observed a trend toward a slightly smaller compressive pressure in the DegUnld group compared to the DegLd group. However, an analysis of individual samples revealed no differences in the biological or biomechanical responses between those that followed this trend and those that did not.

In summary, the compressive pressure achieved during the 50% disc height compression of the one‐strike degenerative trigger on Day 0 was consistent across all groups (Degeneration of Timepoint 1, DegLd and DegUnld of Timepoint 2), ensuring comparability of the experimental conditions.

The accuracy of the applied compression or traction loading during the 2‐h dynamic loading or unloading sessions was evaluated using the root mean square error (RMSE) for both timepoints. The RMSE was calculated for the peak and the valley values of the loading regimes in each group and reported as both a percentage (%) of the target load and as an absolute value (N) (Table 2). In this context, peaks correspond to the minimum compressive force in the loading regime and to the maximum traction force in the unloading regime, whereas the valleys represent the maximum compressive force in both the loading and unloading conditions.

TABLE 2.

Root mean square error (RMSE) of the dynamic sessions in percentage of the target loading (%) and as absolute value (N).

Group	Day 1		Day 2—Day 4
	Loading	Degeneration	Loading	Degeneration + loading	Degeneration + unloading
Percentage target loading (%)
Peak	0.27 (10.5 a )	0.31	2.01	2.16	1.45
Valley	0.04	0.05	0.21	0.04	4.56
Absolute (N)
Peak	0.03 (1.14 a )	0.04	0.05	0.06	0.09
Valley	0.02	0.03	0.11	0.02	0.11

^a Due to an error in the loading command of one sample out of the 9 samples of Timepoint 1_Loading (−1.97 N entered instead of −9.45 N for the first 605 cycles out of 1440), the RMSE is high (10.5%). However, considering this command error, the RMSE of the Loading on Day 1 is good (< 1%). The short loading error didn't seem to affect the results; we therefore decided to keep the sample in the study.

On average, the RMSE was 1% of the target loading, consistent with values previously reported by Walter et al. [39]. However, a slightly lower accuracy (4.56%) was observed for the maximum compressive force in the traction loading regime (i.e., valley in the DegUnld group). This discrepancy is likely due to the technical limitations of the 250 N load cell.

3.2. Timepoint 1: Biological and Biomechanical Impact of the One‐Strike Degenerative Trigger

3.2.1. Biological Evaluation

The NP cell viability (Figure 4A) was significantly lower in the Degeneration group compared to the control Loading group (p < 0.05) with mean values of 87.5% ± 5.5% versus 91.9% ± 6.1% (mean ± standard deviation (SD)). In contrast, the mean viability of the AFi and AFo cells remained above 90% in both groups (cf. Figure S1).

FIGURE 4.

Biological and biomechanical evaluation parameters on Timepoint 1 illustrating the degeneration induced by the one‐strike. Compared to the control Loading group, the nucleus pulposus (NP) samples from the Degeneration group showed, on the biological side, (A) a lower viability (p < 0.05), (B, C) a higher gene expression of MMP3 (p < 0.0001) and iNOS (p < 0.01) and (D) a higher sGAG/water ratio (trend, indicated by the pink arrow, when outlier included and p < 0.05 when outlier excluded). (E) The sGAG release in the medium was also higher in the Degeneration group compared to the loading group (p < 0.0001). Looking at the mechanical evaluation parameters, the Degeneration group presented (F) a lower stiffness of the elastic zone (p < 0.0001), (G) a lower stiffness increase during the loading session (p < 0.05) and (H) a higher hysteresis (p < 0.001). The data are represented as mean ± standard deviation except for (D) where the median is indicated. The log10 transformations are reported on the graphs. If any, the outliers, defined by the Grubbs' test, are indicated with empty symbols (□) and the results of the statistical tests, with and without outliers, are described and specified on the graphs with a gray significance symbol (*). (A, D) In both cases, excluding the statistical outlier resulted in a normal distribution of the data. (A) Excluding the statistical outlier did not influence the level of statistical significance of the paired t‐test. (A) *One of the NP tissue samples lacked in the histological section. (D) *From one sample, we could not collect enough NP tissue to analyze the water/sGAG content in the tissue.

In the NP, the gene expression of MMP3 and iNOS was significantly higher in the Degeneration group than in the Loading group. The fold change difference between groups (95% confidence interval (CI)) of the original and log transformed data was [9.3–57.2 and 1.1–1.7] with p < 0.0001 (Figure 4B) for MMP3 and [−42.1–285.7 and 0.2–1.0] with p < 0.01 (Figure 4C) for iNOS. No significant differences were observed for the other genes analyzed (Figure S2). In the annulus fibrosus, significant differences were detected only in the AFi, where MMP3 (p < 0.001) and TRPV4 (p < 0.05) were elevated in the Degeneration group, suggesting catabolic activity in this region as well (Figures S3 and S4).

A trend towards higher water/sGAG ratio was observed in the NP of the Degeneration group compared to the Loading group (Figure 4D). After exclusion of an outlier (identified by the Grubbs' test), the difference became statistically significant (p < 0.05) with a normal data distribution. Given this, we report the median rather than the mean, as it better represents the trend. The median of the water/sGAG ratio was 44.6 mg/mg in the Loading group and 64.7 mg/mg in the Degeneration group. No significant differences were observed in the annulus fibrosus (Figure S1).

Additionally, sGAG release in the medium was significantly increased in the Degeneration group compared to the Loading group (p < 0.0001), with values of 54.9 ± 16.9 μg versus 27.7 ± 1.7 μg (mean ± SD), respectively (Figure 4E).

3.2.2. Biomechanical Evaluation

The one‐strike degenerative trigger altered the mechanical behavior of the IVD. The stiffness (slope of the elastic zone) at the start of the first loading session (Cycle 30, Day 1) was significantly lower in the Degeneration group compared to the Loading group (p < 0.0001) with mean ± SD values of 202.5 ± 14.8 N/mm versus 241.4 ± 22.0 N/mm, respectively (Figure 4F). Furthermore, the increase in stiffness from Cycle 30 to Cycle 1439 during the Day 1 loading session was significantly lower in the Degeneration group compared to the Loading group (p < 0.05) with 118.2 ± 23.5 N/mm versus 137.9 ± 20.3 N/mm, respectively (Figure 4G).

The hysteresis (area under the curve, AUC) at Cycle 30 on Day 1 was significantly higher in the Degeneration group compared to the Loading group (p < 0.001), with 1.1 ± 0.5 Nmm versus 1.7 ± 0.5 Nmm, respectively (Figure 4H). However, the change in hysteresis over the course of loading did not differ significantly between groups (Figure S1).

In summary, the one‐strike trigger successfully induced a degenerative state in the IVD, as evidenced by both biological and biomechanical disruptions. Disrupted homeostasis in the NP was demonstrated by a reduced viability, increased catabolic gene expression, and elevated sGAG release in the medium (Figure 4A–C,E). Altered biomechanical response was characterized by reduced stiffness and increased hysteresis at the start of the loading, along with a smaller stiffness increase during the loading session (Figure 4F–H).

3.3. Timepoint 2: Impact of Loading and Unloading on Degenerated IVDs

On Timepoint 2 (Day 4), we evaluated the effects of unloading on degenerated IVDs through biological and mechanical parameters. Specifically, we compared unloading following the one‐strike trigger to continued loading after the one‐strike and to loading alone (without degeneration). Since the degenerative trigger primarily affected the NP on Timepoint 1, we focused on Timepoint 2 on significant NP‐related findings. Additional results from Timepoint 2 are provided in the Figures S5–S8.

3.3.1. Biological Evaluation

3.3.1.1. Cell Viability

The cell viability in the DegLd and DegUnld groups was lower than in the Loading control group (72.0% ± 13.1% vs. 68.4% ± 18.8% vs. 79.1% ± 19.9%, respectively). After excluding one outlier in the Loading group (Grubb's test) the DegUnld exhibited significantly lower viability compared to the Loading group (p < 0.05) (Figure 5). Additionally, six out of eight DegLd samples had lower viability than their corresponding Loading group samples. Across all groups, NP cell viability decreased by 10%–15% from Timepoint 1 to Timepoint 2, reflecting the influence of culture conditions on cell viability (Figures 4A and 5).

FIGURE 5.

Viability of the nucleus pulposus on Timepoint 2 suggested less cells alive in the two degeneration groups. Compared to the control group Loading, the nucleus pulposus (NP) samples from the DegLd and DegUnld group showed a trend, indicated by the pink arrow, towards a lower viability (with Loading >< DegUnld p < 0.05 when outlier excluded). The data are represented as mean ± standard deviation. The outlier, defined by the Grubbs' test, is indicated with an empty symbol (□) and the results of the statistical tests, with and without outlier, are described and specified on the graphs with a gray significance symbol (*). The exclusion of the statistical outlier resulted in a normal distribution of the data. *One of the NP tissue samples lacked in the histological section.

To conclude, on Timepoint 2, NP cell viability tended to be lower in both DegLd and DegUnld groups compared to the Loading control group, suggesting continued degenerative effects.

3.3.1.2. Gene Expression

COL1A1 expression was significantly upregulated in both DegLd and DegUnld groups compared to the Loading group, which did not receive a degenerative trigger (p < 0.0001 for both comparisons; Figure 6A). MMP3 expression was also significantly higher in the DegLd and DegUnld groups than in the Loading group (p < 0.0001 for both comparisons; Figure 6B).

FIGURE 6.

Gene expression of COL1A1, MMP3, iNOS, TRPV4, and AQP1 in the nucleus pulposus (NP) on Timepoint 2 indicated a trend towards a decreased inflammatory and osmosensing gene expression in the DegUnld group. (A) COL1A1 was higher expressed in the DegLd and DegUnld groups compared to the Loading group (p < 0.0001 and p < 0.0001). (B) MMP3 was higher expressed in the DegLd and DegUnld groups compared to the Loading group (p < 0.0001 and p < 0.0001). (C) iNOS was also higher expressed in the DegLd and DegUnld groups compared to the Loading group with a greater difference between the DegLd and Loading groups than between the DegUnld and Loading groups (p < 0.01 and p < 0.05 respectively). (D) TRPV4 was only higher expressed in the DegLd group compared to the Loading group. (E) AQP1 was less expressed in the DegLd and DegUnld groups compared to the Loading group (p < 0.05 and p < 0.01 respectively). The data are represented as mean ± standard deviation. The pink arrows indicate trends. All gene expression data were log10 transformed.

iNOS expression was significantly increased in both DegLd and DegUnld groups compared to the Loading group, but the upregulation was greater in the DegLd group (p < 0.01 for DegLd vs. Loading; p < 0.05 for DegUnld vs. Loading; Figure 6C). A trend towards lower iNOS expression was observed in the DegUnld group compared to the DegLd group. TRPV4 was significantly upregulated in the DegLd group compared to the Loading group (p < 0.05; Figure 6D). However, a trend towards lower TRPV4 expression was seen in the DegUnld group compared to the DegLd group. AQP1 expression was significantly lower in both DegLd and DegUnld groups compared to the Loading group (p < 0.05 and p < 0.01, respectively; Figure 6E).

The significant differences between the groups expressed in fold changes (95% confidence interval (CI)) of the original and log transformed data are detailed in Table 3.

TABLE 3.

Difference in fold changes between the groups: 95% confidence interval and p value.

Gene		Loading versus DegLd		Loading versus DegUnld
COL1A1	Original data	−8.0 to 3.9	p < 0.0001	−10.1 to 1.8	p < 0.0001
	Log transformed data	−1.3 to −0.6		−1.4 to −0.6
MMP3	Original data	−27.3 to −5.3	p < 0.0001	−20.9 to 1.1	p < 0.0001
	Log transformed data	−1.9 to −0.9		−1.9 to −0.8
iNOS	Original data	−34.8 to 1.4	p < 0.01	−32.9 to 3.3	p < 0.05
	Log transformed data	−1.0 to −0.2		−0.8 to −0.05
TRPV4	Original data	−1.1 to 0.1	p < 0.05	−0.0 to 0.3	NS
	Log transformed data	−0.6 to −0.02		−0.5 to 0.06
AQP1	Original data	−0.3 to 2.0	p < 0.05	−0.3 to 1.9	p < 0.01
	Log transformed data	−0.03 to 0.5		0.1 to 0.6

To conclude, gene expression analysis revealed persistent degeneration in both DegLd and DegUnld groups, with sustained catabolic activity (COL1A1, MMP3). However, the trends toward reduced iNOS and TRPV4 expression in the DegUnld group compared to the DegLd group suggest that unloading may partially reduce inflammation and osmotic stress.

3.3.1.3. Water‐To‐sGAG Ratio and sGAG Release

No statistical differences in the water/sGAG ratio were observed between the groups (Figure 7A). However, there was a trend towards a lower ratio in the DegLd group compared to the Loading group (6 out of 9 samples; 45.3 ± 14.9 vs. 60.6 ± 29.1). In contrast, the DegUnld group showed a trend towards a higher ratio compared to DegLd (7 out of 9 samples; 62.7 ± 29.1 vs. 45.3 ± 14.9), reaching levels similar to the Loading group.

FIGURE 7.

Water/sGAG ratio in the nucleus pulposus (NP) and sGAG release in the medium on Timepoint 2 illustrated in the DegUnld group a trend towards recovery of the water/sGAG ratio to the ratio of the Loading group and a trend towards a decrease in sGAG release in the medium compared to the DegLd group. (A) We observed a trend towards a decrease of the water/sGAG ratio in the DegLd group compared to the Loading and DegUnld groups. The outlier, defined by the Grubbs' test, is indicated with an empty symbol (○). The exclusion of the statistical outlier resulted in a normal distribution of the data but did not influence the results of the ANOVA. (B) The sGAG release in the medium increased in the DegLd group (p < 0.01) compared to the Loading group. We also observed a trend towards a decrease in sGAG release in the medium of the DegUnld group compared to the DegLd group. (A, B) The data are represented as mean ± standard deviation. The pink arrows indicate trends.

sGAG release was significantly increased in the DegLd group compared to the Loading group (p < 0.01; 81.5 ± 21.4 μg vs. 47.2 ± 22.7 μg, respectively; Figure 7B). A trend toward reduced sGAG release was observed in the DegUnld group compared to DegLd (7 out of 9 samples; 66.3 ± 26.3 μg vs. 81.5 ± 21.4 μg).

To conclude, on Timepoint 2, the water/sGAG ratio in the DegUnld group resembled that of the Loading group, suggesting partial recovery of proteoglycan function. Similarly, the trend towards reduced sGAG release in the DegUnld group may indicate a shift towards proteoglycan retention.

3.3.1.4. Morphological and Proteoglycan Degradation Analyses

Safranin O/Fast Green staining revealed no visible differences between the groups (Figure S9).

In the agarose gel electrophoresis evaluation, no evidence of proteoglycan degradation was detected, and no differences were observed between the groups (Figure S10).

These results confirm the preservation of the extracellular matrix during early‐stage degeneration, consistent with previous studies.

3.3.2. Biomechanical Evaluation

3.3.2.1. Disc Height

Following the one‐strike impact, the disc height decreased by ~20% after dissection (Figure 8). Compared to the nondegenerated Loading group, the reduced disc height persisted in the DegLd group throughout the culture period (overall mean ± SD: 11.3 ± 0.9 mm vs. 10.2 ± 0.8 mm). In the Loading and DegLd groups, the disc height decreased during dynamic loading sessions but recovered during static loading. In contrast, in the DegUnld group, the disc height increased during dynamic unloading sessions but decreased during static loading. From Day 2 on to Day 4, disc height in the DegUnld group had recovered to values similar to the Loading group (p > 0.05; overall mean ± SD: 11.4 ± 1.3 mm vs. 11.3 ± 0.9 mm).

FIGURE 8.

Relative (mean ± standard deviation) and absolute disc height change over time showed a decrease in disc height following the one‐strike that was maintained over time and a recovery of the disc height in the DegUnld group to the disc height of the Loading group. (A, B) The one‐strike degenerative trigger decreased the disc height. Globally, the reduced disc height in the DegLd group was maintained over time. In the Loading and DegLd groups, the disc height decreased during the dynamic sessions but increased during the static loading, in contrast to what was observed in the DegUnld group. Additionally, the DegUnld group showed increasing disc height over time, in contrast to the other two groups (p < 0.0001 and p < 0.0001), and the DegUnld group had a disc height similar to the Loading group.

Over time (Day 2 to Day 4), disc height changes differed significantly between the Loading and DegLd groups compared to DegUnld (p < 0.0001 for both comparisons). Although the disc height declined in the Loading and DegLd groups (slope: −0.10 ± 0.02 and −0.12 ± 0.04, respectively), the DegUnld group exhibited an increasing trend (slope: 0.06 ± 0.04).

To conclude, the one‐strike impact caused a sustained disc height reduction in the DegLd group, whereas unloading reversed this trend, leading to disc height recovery and an increasing disc height over time.

3.3.2.2. Stiffness and Hysteresis

Stiffness (elastic zone slope) was significantly lower in the DegLd group compared to the Loading group at all measurement points (p < 0.0001; Figure 9A).

FIGURE 9.

Evolution of the stiffness of the elastic zone and the hysteresis over time illustrated a persistent lower stiffness and higher hysteresis over time in the DegLd group. (A) The stiffness increased during the dynamic sessions (Cycles 30 to 1439) and decreased during the static loading (Cycles 1439 to 30). The three groups exhibited a similar pattern over time. (B) The hysteresis showed the opposite pattern. (A, B) The impact of the one‐strike degenerative trigger, expressed as a lower stiffness and higher hysteresis in the DegLd group compared to the Loading group, persisted over time (p < 0.0001 on every measurement point). The data are represented as mean ± standard deviation.

Hysteresis (AUC) remained higher in the DegLd group compared to the Loading group at all measurement points (p < 0.0001; Figure 9B).

Although stiffness and hysteresis changed dynamically within each session, the overall trend remained unchanged, indicating that the detrimental effects of the one‐strike persisted over time in the DegLd group.

To conclude, the stiffness loss and increased hysteresis observed after the one‐strike impact persisted in the DegLd group, reinforcing the notion that degeneration was maintained over the culture period.

3.3.2.3. Neutral Zone

Displacement range of the neutral zone decreased over time (slope: −0.02 ± 0.02) and during dynamic unloading sessions, whereas it increased during static loading (Figure 10).

FIGURE 10.

Evolution over time of the different characteristics of the neutral zone (force, displacement and stiffness) in the DegUnld group showed a decrease in the displacement during the unloading sessions and over time. The displacement range of the neutral zone decreased while the force range of the neutral zone increased during the unloading sessions and over time. The data are represented as mean ± standard deviation. Despite the high standard deviation, the mean is representative of the individual data points.

Force range increased over time (slope: 0.02 ± 0.12).

The progressive reduction in neutral zone displacement suggests that unloading may help stabilize degenerative‐associated instability.

4. Discussion

In this bovine tail IVDs bioreactor study, we investigated the biological and biomechanical impact of a one‐strike degenerative trigger on Timepoint 1 and the subsequent impact of short‐term dynamic unloading compared to loading on Timepoint 2. The biological changes were assessed through cell viability, gene expression, water/sGAG ratio, and sGAG release in the medium, whereas mechanical modifications were evaluated by measuring disc height, elastic zone slope, area under the curve (AUC) and neutral zone characteristics. This study provides an initial insight into the mechanobiology of unloading in degenerated IVDs and highlights the interplay between biological and biomechanical responses in the IVD. Our findings suggest that short‐term dynamic unloading may partially mitigate the degenerative effects in the NP, emphasizing the need for further research to better understand its regenerative potential.

The one‐strike degenerative trigger effectively induced a degenerative state in the NP on Timepoint 1. This was evidenced by a reduction in cell viability, suggesting acute cell death, and a significant increase in MMP 3 (matrix degradation enzyme) and iNOS (inflammatory marker) gene expression, indicating a shift towards a catabolic phenotype. Furthermore, the increased sGAG release in the Degeneration group points to proteoglycan loss, a hallmark of disc degeneration. Biomechanically, degeneration was reflected by reduced stiffness and increased hysteresis at the start of the dynamic loading session, indicating a loss of structural integrity and mechanical efficiency. Additionally, the slower increase in stiffness over time in the Degeneration group suggests that the mechanical resilience of the disc was compromised following impact loading. Overall, six key biological and biomechanical parameters exhibited significant differences between the Degeneration and Loading groups, confirming the degenerative [7] changes induced by the one‐strike impact.

These findings are consistent with previous studies on human and bovine IVDs, which have reported similar changes in viability, matrix composition, and gene expression following mechanical injury [40, 41]. Interestingly, the increased water‐to‐sGAG ratio in the NP of the Degeneration group, combined with the lack of significant hysteresis change during dynamic loading, suggests that the one‐strike trigger not only initiates degeneration but also induces a compensatory increase in water retention on Timepoint 1. This reactional water intake may represent an attempted repair response, as water plays a critical role in maintaining intradiscal pressure and overall disc function [11, 14]. However, due to the study design, it was not possible to differentiate the intrinsic repair response of the IVD from the effects of subsequent loading regimes on Timepoint 2. Future studies should focus on quantifying the interaction between degenerative triggers and loading regimes, as well as investigating the long‐term effects of acute mechanical injury and the disc's intrinsic repair capacity.

On Timepoint 2, the degenerative cascade persisted in both degeneration groups (DegLd and DegUnld). However, unloading exhibited trends suggesting a partial mitigation of degeneration, particularly in the NP.

In the DegLd group, which underwent continued loading, the decrease in cell viability, increased expression of COL1A1, MMP3, iNOS, and TRPV4, decreased expression of AQP1 (water channel protein), increased sGAG release, and maintained reductions in stiffness and disc height all indicate that compression loading of a degenerated disc maintains a compromised homeostasis [7, 8, 10, 15, 42]. Conversely, in the DegUnld group, although COL1A1 and MMP3 expression remained elevated, indicating continued degeneration, there were notable trends suggesting a partial recovery. Specifically, the downward trend in iNOS and TRPV4 expression compared to the DegLd group suggests a reduction in inflammation and a shift towards increased osmolarity, resembling the gene expression profile of the undegenerated Loading group. These findings align with previous research demonstrating the beneficial effects of unloading in undegenerated IVD [28]. The lower sGAG release in the DegUnld group, positioned between the DegLd and Loading groups, suggests a tendency for proteoglycan retention in unloaded IVDs. Additionally, the water‐to‐sGAG ratio in the DegUnld group, which was comparable to that of the Loading group, points to a partial restoration of functional proteoglycans and an increase in NP hydration. The recovery of disc height in the DegUnld group, reaching values similar to those in the undegenerated Loading group, further supports this interpretation.

Increases in disc height and water content following unloading protocols have been reported in animal studies and clinical settings [21, 22, 23, 43, 44, 45, 46, 47]. Additionally, the trend towards reduced AQP1 expression in the DegUnld group may not necessarily indicate a degenerative process, but rather an adaptive response to enhanced water flow, demonstrated by Kuo et al. [47] in porcine IVDs. Since unloading appears to increase water influx, the need for AQP1‐mediated transport might be reduced.

The progressive decrease in the displacement range of the neutral zone during unloading suggests that unloading, combined with the associated biological changes, may eventually reduce degeneration‐associated instability [14, 15, 16]. This underscores the interdependence of biological and mechanical factors in the IVD and highlights the potential therapeutic benefits of unloading in disc degeneration.

The observed trends of partial recovery in the biological and biomechanical parameters suggest that unloading is a promising area for further investigation. The continuous increase in disc height over time in the DegUnld group, combined with the absence of adverse morphological changes after degeneration, supports the necessity of longer‐term bioreactor studies to explore the full effects of unloading on degenerative IVDs, concordantly with previous unloading studies in small animals [21, 22, 23]. Additionally, different unloading amplitudes, similar to those used in clinical traction [48, 49], should be examined. Future research could also evaluate other exercise‐based or therapeutic movements that promote water retention in the IVD [50, 51].

Several limitations should be acknowledged. Firstly, the limited timeframe of the culture restricts the evaluation of long‐term matrix remodeling following degeneration and unloading. However, as this was a pilot study, a shorter duration was justified. Secondly, the presence of unsystematic endplate fractures following the one‐strike impact could have influenced the results. Nonetheless, the pressure applied was consistent across samples, and no outliers were detected due to endplate fractures. Thirdly, the high variability in certain biological and mechanical parameters limited the detection of statistically significant differences. A larger sample size could help clarify trends and increase statistical power. However, practical constraints of the current experimental setting and the typically small sample sizes used in this research field posed challenges. Finally, the induced degeneration primarily affected the NP, whereas more advanced or widespread degeneration was not modeled. Future studies should explore unloading in different degeneration models, including chronic, more progressive, or chemically induced degeneration. Also, longer‐term studies should be designed to assess the effect of dynamic unloading on extracellular matrix remodeling. Finally, valuable insight into the response of induced degenerative IVDs to dynamic unloading can be gained from bovine specimens that resemble human IVDs in terms of size, cell and matrix composition. Nevertheless, future work should replace bovine with human tissues to facilitate the translation from the laboratory to the patient.

5. Conclusion

This study demonstrates that a one‐strike mechanical degenerative trigger successfully induces a NP degenerative cascade on Timepoint 1. On Timepoint 2, degeneration was maintained in both degeneration groups; however, short‐term dynamic unloading showed trends indicating a partial recovery of biological and biomechanical homeostasis compared to continued loading.

This first investigation into the unloading mechanobiology of degenerated bovine IVDs in culture provides promising preliminary findings, supporting the exploration of exercise and therapeutic movement as regenerative strategies for IVD degeneration. Further research should focus on longer culture durations, different degenerative models, multiple unloading regimes, and molecular transport assessments before progressing to translational and clinical studies.

Author Contributions

Astrid Soubrier: writing – original draft, conceptualization, methodology, investigation, validation, formal analysis, data curation, visualization. Hermann Kasper: methodology, conceptualization. Nadja Vonlanthen: data curation, investigation, visualization. Ilse Jonkers: conceptualization, writing – review and editing, supervision, methodology. Sibylle Grad: conceptualization, methodology, supervision, project administration, writing – review and editing, resources, funding acquisition.

Disclosure

Sibylle Grad is an Editorial Board member of JOR Spine; she is the corresponding author of this article. She was excluded from editorial decision‐making related to the acceptance of this article for publication in the journal.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1. jsp270092‐sup‐0001‐supinfo.

Soubrier A., Kasper H., Vonlanthen N., Jonkers I., and Grad S., “Short‐Term Dynamic Unloading of Bovine Tail Discs in Culture Partially Mitigates Induced Degeneration After One‐Strike Trigger,” JOR Spine 8, no. 3 (2025): e70092, 10.1002/jsp2.70092.

Data Availability Statement

The data that supports the findings of this study are available in the Supporting Information of this article.