Injury Induces More Severe Biomechanical Changes in Middle‐Aged and Geriatric Lumbar Spines in a Mouse Ex Vivo Model

ABSTRACT

Background

Intervertebral disc degeneration (IVDD) is a major cause of global disability that increases with age. IVD age may affect its injury susceptibility, yet few studies examine spine biomechanical changes with age, fewer address multiple injury types, and none investigate the interplay between age and injury.

Methods

An ex vivo mouse lumbar spine biomechanical study to determine the effects of age, injury, and their interaction. IVDs of 4, 12, and 24 months' mice were subjected to two injury types: Full disc puncture (DP) mimicking advanced IVDD and annulus fibrosus and endplate (AF + EP) injury simulating herniation with endplate junction failure. Spines were tested biomechanically, analyzed radiologically for IVD dimensions, and with FTIR and histology for biochemical content.

Results

Both age and injury significantly altered biomechanical properties of IVDs. Injury had a greater effect than age, and DP caused larger changes than AF + EP injury. Injury and age exhibited an interactive effect, resulting in more pronounced biomechanical dysfunction in middle‐aged (12 months) and geriatric IVDs (24 months), likely due to age‐related loss of proteoglycans and collagen denaturation shown with FTIR and histology.

Conclusions

We conclude that both age and injury independently and synergistically affect ex vivo biomechanical properties of mouse lumbar spine. The more severe biomechanical change in middle‐aged and geriatric mouse lumbar spines suggests similar injuries may cause greater spinal dysfunction in individuals of comparable ages. These findings provide context for future in vivo studies investigating long‐term effects of acute spine injuries.

Injury and age exhibited an interactive effect, resulting in greater biomechanical dysfunction in middle‐aged (12 months) and geriatric (24 months) mouse intervertebral discs (IVDs), associated with age‐related proteoglycan loss and collagen denaturation as shown by FTIR and histological analyses.

1. Introduction

Low back pain is a common condition experienced by most individuals at some point in their lifetime, with its prevalence increasing with age and being strongly associated with functional impairment [1, 2]. Back pain is a significant concern, especially in the older population, with self‐reported prevalence of lower back pain being highest among middle‐aged and elderly individuals [3, 4, 5]. In a geriatric population, chronic lumbar pain subjects had reduced flexion and extension range of motion compared to non‐chronic lumbar back pain subjects, suggesting age‐related biomechanical changes in the spine [2]. Further, by 2050, it is projected that 89 million individuals in the USA will be aged 65 and older, highlighting a need for studies on age‐related spinal disability [6].

Spinal biomechanical dysfunction, a contributing factor to back pain, can result from various pathological IVD disorders associated with an aging spine, such as IVDD or herniation [7, 8, 9], endplate (EP) damage [10, 11], pathological remodeling after injury [12, 13], and abnormal biomechanical loading [14, 15, 16]. In the lumbar spine, IVDs are the primary structures implicated in functional impairments and pain. IVDs are comprised of a central gelatinous nucleus pulposus (NP), containing water, collagen II, and proteoglycans; surrounded by concentric layers of collagen‐rich annulus fibrosus (AF). The NP maintains IVD height and hydration through osmotic pressurization, while the AF contains NP pressurization, connects the vertebrae, and resists tensile forces during bending and twisting. The NP pressure is constrained on the superior and inferior surfaces by cartilage and vertebral EPs and is evenly distributed against the EP surface. The EP prevents the NP from bulging into the trabecular bone of the vertebrae [17, 18].

Aging is associated with loss of IVD height, water, glycosaminoglycan (GAG) content, and increased collagen denaturation [19, 20, 21, 22, 23, 24, 25]. Loss of water and GAG contents shift load carrying mechanisms and lead to reduced pressurization and increased AF matrix strains, putting the IVD at risk of progressive damage and IVDD [9, 18]. A previous study showed these age‐related compositional changes reduced tensile stiffness and viscoelastic coefficient in mouse IVD motion segments [26]. Despite the well‐known changes in the IVD with aging, we found no studies addressing the interaction of age and injury on lumbar IVD in pre‐clinical studies, and we believe both factors could interact to cause a greater biomechanical dysfunction from injury in the aged‐IVD.

We used a mouse model for this study because mice have a substantially shorter lifespan and lower variability compared to humans. Mouse lumbar IVDs closely resemble human IVD anatomy when scaled for size, yet there are few studies on mouse lumbar spine biomechanics due to challenges in accessing the lumbar spine and low post‐surgical survival rates in mice [26, 27, 28, 29]. Despite these challenges, the benefits of mouse models have motivated many mouse IVDD studies investigating mechanical property changes in the IVD. Mouse tail IVD models are commonly used because of the easy access and simplified surgical approach compared to lumbar IVDs in mice [30, 31, 32, 33, 34, 35]. However, there are known structural, compositional, micro‐environmental, and biomechanical differences between tail and lumbar IVDs, and mouse lumbar IVDs have an anatomy that is more relevant to the investigation of human lumbar IVDD and back pain conditions [26, 36, 37].

This study determined the interacting effects of age and IVD injury on acute spine biomechanical properties using a mouse lumbar spine ex vivo model. It was hypothesized that age and injury would both affect IVD biomechanical properties and would interact to affect axial and creep properties due to NP depressurization and torsional properties due to AF damage. We also hypothesized that greater injury severity would induce more substantial biomechanical changes. Therefore, two injury conditions were applied to mouse lumbar spine motion segments of three ages, which represent normal health at young‐adult (4 months), middle‐age (12 months), and geriatric (24 months) [24]. We highlight that this is an acute ex vivo injury study without attempted healing or progressive degeneration. The AF + EP injury involved a posterolateral needle puncture of the AF at the EP junction of the annular enthesis, to simulate an acute IVD herniation with endplate junction failure [38]. The full disc puncture (DP) injury consisted of a complete bilateral transannular puncture, simulating structural disruption of AF and NP compartments in IVDD [35].

2. Materials and Methods

2.1. Animal Information and Lumbar IVD Injury Type

With IACUC approval, lumbar motion segments (L1‐2, L3‐4, L5‐6) were isolated from mice of three age groups (within 1 week of): 4 months (young‐adult), 12 months (middle‐aged), and 24 months (geriatric) with n = 7–9 (3–4 males and 4–5 females) per group (Figure 1). Motion segments from each animal were assigned either Intact, DP, or AF with EP defect injury (AF + EP), with level systematically rotated across each group to account for potential level effects (Figure 1). DP was an injury involving a 26G needle creating a midline puncture through the anterior AF, passing through the NP, and exiting at the posterior AF periphery. AF + EP was an injury involving a 26G needle injury from anterior AF obliquely punctured through the superior EP to simulate an IVD herniation. X‐ray (Faxitron) images confirmed injury in all specimens.

FIGURE 1.

Schematics showing experimental setup, IVD injury types, and outcomes. IVDs from different spinal levels (L1‐2, L3‐4, L5‐6) were analyzed to assess age‐related changes in IVD properties. Structural properties were evaluated using Faxitron imaging, biochemical composition of L5‐6 spinal IVD through histological analysis, and biophysical properties (L5‐L6) using FTIR. Different spinal levels were designated as either internal controls (Intact) or subjected to injury models, including full disc puncture (DP) mimicking advanced IVDD and combined annulus fibrosus and endplate (AF + EP) injury simulating IVD herniation with endplate junction failure. Biomechanical testing was performed on both Intact and injured IVDs, including axial cyclic loading (20 cycles of ±0.8 N tension/compression at 1 Hz), axial creep (rapid application of −0.8 N compression, held for 45 min), and torsional testing (20 cycles of ±20° angular rotation at 1 Hz). The IVD was rehydrated in PBS for 45 min between creep and torsional testing.

2.2. Biomechanical Testing

Axial tension‐compression and creep tests were conducted using an Electroforce 3200 instrument (TA Instruments, Eden Prairie, MN, USA) under a force‐controlled cyclic testing protocol. Spine motion segments (Intact, DP, AF + EP), comprising vertebra‐IVD‐vertebra units, were potted between two cylindrical metal fixtures using cyanoacrylate glue and immersed in a phosphate‐buffered saline (PBS) bath at room temperature throughout the testing procedure. The axis of the motion segment was precisely aligned with the axis of the two parallel fixtures to ensure proper load transfer, and the center of the IVD was positioned equidistantly from the edges of the fixtures on both sides (Figure 1). Axial cyclic loading tests were performed by applying 20 cycles of ±0.8 N (~0.4 MPa) tension/compression at a frequency of 1 Hz (Figure 1). Following the cyclic testing, the samples were allowed to rest for 4 min, after which the viscoelastic properties of the IVD were evaluated through axial creep testing. This involved rapidly applying a constant compressive force of 0.8 N, which was maintained for 45 min (Figure 1). Subsequently, the motion segments were rehydrated in PBS‐soaked gauze for 45 min before cyclic torsion testing. Axial tension‐compression properties, including compressive stiffness, tensile stiffness, range of motion (ROM), neutral zone (NZ) length, NZ stiffness, and hysteresis, were calculated from the 19th cycle using MATLAB (MathWorks Inc., Natick, USA). Creep data were fitted into a five‐parameter exponential model to characterize the viscoelastic behavior of the IVD, enabling the calculation of fast and slow time constants and displacement losses (fast, slow, elastic) across different creep phases [39]. In this five‐parameter model: the elastic constant is related to the immediate displacement of the structure; the fast time constant and displacement are predominantly related to flow‐independent viscoelasticity due to frictional loss between solid components of the tissues including radial bulging under load; and the slow time constant and displacement are related to flow‐dependent, or biphasic viscoelastic behaviors.

Torsional testing was performed on an AR2000‐EX rheometer (TA Instruments, New Castle, DE, USA) using an angular displacement‐controlled protocol. The potted motion segment was fixed in the rheometer, with rotational motion applied from one end while the other end remained stationary to provide rotational movement to the IVD. A humidifying chamber created using water‐soaked gauze on a 37°C warming plate within a small plastic container kept the IVD hydrated (Figure 1). A compressive preload of −0.5 N was applied for 5 min to simulate physiological loading and ensure consistent testing. After preloading, 20 cycles of ±20° angular rotation at 1 Hz were applied (Figure 1). The 20° range falls within the axial rotation, lateral bending, and flexion/extension range typical of human daily activities [40]. Torsional properties, including torque range, stiffness, NZ length, and hysteresis, were calculated from both the 3rd (early cycle) and 19th (late cycle) cycles using MATLAB (MathWorks Inc., Natick, USA) code and compared to assess differences between early and late cycle data. The 3rd cycle is the first full cycle after a single pre‐conditioning cycle.

2.3. IVD Structural Properties

IVD height was measured by averaging the anterior, posterior, and center heights to obtain the mean IVD height for all three age groups. The IVD area was calculated by measuring the lateral width (w1) and anteroposterior width (w2) and applying the formula for the area of an ellipse: Area = (3.14 * w1 * w2)/4 [27]. IVD circularity was measured by dividing the anteroposterior width (w2) by the lateral width (w1), and convexity was calculated by dividing the center IVD height by the average of the anterior and posterior heights [41, 42]. These measurements were taken using an X‐ray and a digital caliper.

2.4. Fourier Transform Infrared Spectroscopy (FTIR)

FTIR is a non‐invasive tool used to measure tissue biochemical composition like collagen and proteoglycan [43, 44]. FTIR imaging using transmittance mode or powdered samples mixed with KBr pellets shares the same underlying physical chemical principle as the ATR mode. As a semi‐quantitative comparison, the relative contribution of analyzed matrix components in ATR would match the FTIR analysis in other modes. Infrared spectra were obtained using a Bruker Vertex 70 spectrometer in attenuated total reflectance mode (Figure 1). L5‐L6 IVDs (n = 6/age) were pad‐dried and then air‐dried for 1 h to remove the surface free water, maximizing the OH stretch signal that is bound to the matrix, and in ATR mode, the IVD is centered and pressed against a 1 mm × 1 mm diamond, whose measured area covers the entire inferior IVD surface and is scanned 256 accumulations with a 2 cm⁻¹ spectral resolution. Peak analysis, including deconvolution and integration, was done in OriginPro software (OriginLab Corporation, Northampton, USA). For each peak of interest, a local linear baseline correction was applied between the valleys of the integrating range to minimize the baseline influence. Peaks for collagen and proteoglycan contents were selected using methods described for cartilage and IVD [43, 44]. Bound water analysis applied peaks determined on bone tissues, and we note the peak attributed to bound water is not influenced by mineral content and closely matches in shape and position with that observed in IVDs [45]. Specifically, proteoglycan content was measured by integrating the 1140–1180 cm⁻¹ range. Organic matrix components including Amide I (1585–1715 cm⁻¹), Amide II (1480–1585 cm⁻¹), CH₂ side‐chain vibrations (1300–1356 cm⁻¹) were quantified by integration of the respective areas. Bound water content was analyzed in the 2500–3700 cm⁻¹ range, with sub‐peaks identified through the second derivative and smoothed using a Savitzky–Golay filter. The sub‐peak at ~3200 cm⁻¹ was integrated to estimate bound water content. The ratios of the sub‐peak to the Amide I peak and the proteoglycan peak to the Amide I peak were used to calculate relative bound water and proteoglycan content. The ratio of CH₂ at ~1388 cm⁻¹ and Amide II was used as an indicator of collagen integrity [46, 47]. The sub‐peak ratio within the Amide I band (1660/1690 cm⁻¹) was quantified as a measure of collagen enzymatic crosslinks [48].

2.5. Histology and Tinctorial Staining of IVD

Following euthanasia, the lumbar spine segment L5‐6 was dissected from 4 month (n = 2 female, n = 3 male), 12 month (n = 3 female, n = 3 male), and 24 month (n = 3 female, n = 3 male) mice in the FVB/NJ background. The spine was immediately fixed in buffered 4% paraformaldehyde (PFA) for 8 h on a rocker platform at 4°C. Following three washes in cold PBS for 30 min each, the spines were decalcified using 0.5 M ethylenediaminetetraacetic acid (EDTA, Sigma‐Aldrich, USA, E9884), pH 7.4 at 8°C on a rocker platform for 7 to 9 days. Next, the spines were washed thrice in cold PBS for 30 min each and individually molded for cryosectioning using Tissue‐Tek optimum cutting temperature (O.C.T, VWR, USA, 102094‐106), snap frozen, and stored at −80°C until further use. Cryosections were prepared in the coronal plane at 8 μm thickness using a Leica cryostat. Slides were stored at −80°C until further use. Mid‐coronal sections were processed for histochemical evaluation.

For Safranin O and Fast Green Staining (SafO/FG), frozen sections were fixed in 4% PFA for 5 min, washed in PBS, and stained using the Safranin O Cartilage Stain Kit—MasterTech (SKU: KTSFOPT, Statlab, McKinney, USA) following the manufacturer's instructions. Next, the slides were dehydrated using an increasing gradient of ethanol, cleared in xylene, and mounted in PROTOCOL Mounting Medium (C.A.S. 23‐245‐691, Fisher Healthcare, Hampton, USA). L5‐6 IVDs were imaged at 10× and 40× magnification under bright field using a Nikon Eclipse wide‐field microscope and accompanying NIS Elements AR software (Nikon, Tokyo, Japan). For Picrosirius Red staining, frozen sections were fixed in 4% PFA for 5 min, washed in PBS, and stained using the Picrosirius Red Stain Kit (Catalog no. 24901‐250, PolySciences INC, Warrington, USA) as per the manufacturer's instructions for frozen sections. Following staining, the slides were dehydrated, cleared, and mounted. L5‐6 IVDs were imaged at 10× and 40× magnifications and crossed polars using a Nikon microscope and accompanying software (Figure 1).

2.6. Statistical Analysis

Using one‐way ANOVA, structural properties (IVD height, area, circularity, and convexity index) and FTIR measurements (proteoglycan and collagen content) were evaluated using GraphPad Prism 10 (GraphPad, San Diego, USA). If significant, Tukey's post hoc test was run to compare groups. For biomechanical properties, a two‐way ANOVA was performed to assess the effects of age, injury, and their interaction (Age × Injury), applying Tukey's post hoc test when significant main effects or interactions were identified. For all analyses, normality was confirmed with the Shapiro–Wilk test prior to performing subsequent parametric statistical tests, and significant differences were detected with p < 0.05.

3. Results

3.1. Aged IVDs Showed a Greater Decrease in Compressive Stiffness After Injury Than Younger IVDs

The axial force‐displacement curve was analyzed to calculate tension‐compression properties, including compressive stiffness, tensile stiffness, ROM, NZ length, NZ stiffness, and hysteresis (Figure 2A). The gray line represented Intact IVDs, the red line denoted DP injuries, and the blue line depicted AF + EP injuries in 4 months (Figure 2B), 12 months (Figure 2C), and 24 months (Figure 2D) IVDs. Age did not significantly affect compressive stiffness (Figure 2E). Injuries significantly reduced compressive stiffness in 12 and 24 months IVDs. In the DP injuries, compressive stiffness decreased by 43% at 12 months and 48% at 24 months compared to Intact IVDs. Similarly, in the AF + EP group, reductions of 43% at 12 months and 41% at 24 months were observed compared to Intact. Age reduced tensile stiffness in 12 and 24 months Intact IVDs compared to 4 months Intact IVDs. Injuries further reduced tensile stiffness, most notably in 4 months DP‐injured IVDs, though no significant differences with injury were observed in the 12 and 24 months age groups (Figure 2F). Age led to a decrease in the ROM most significantly in 24 months IVDs compared to 12 and 4 months IVDs (Figure 2G). Conversely, injuries increased ROM, with a 70% increase observed in 12‐month DP‐injured IVDs (Figure 2G). Both age and injuries increased hysteresis, with the largest changes in 24 months DP‐injured IVDs compared to Intact IVDs. Though age‐injury interaction was trending (p = 0.08), most hysteresis changes were driven by aging, especially in DP groups (Figure 2H). Injury significantly increased the NZ length, particularly in DP‐injured IVDs, with no significant changes in AF + EP injuries (Figure S1A). Age did not affect NZ stiffness, but injuries reduced it, with greater reductions in DP‐injured IVDs compared to Intact at 4 and 12 months IVDs (Figure S1B). In summary, age and injuries independently and interactively influenced axial IVD properties, affecting stiffness, ROM, NZ length, NZ stiffness, and hysteresis in an age‐ and injury‐dependent manner.

FIGURE 2.

Injury caused a greater reduction in axial compressive stiffness in 12 and 24 month IVDs than in 4 month IVDs. (A) Representative force‐displacement curve showing the calculated axial tension‐compression parameters. (B–D) Representative force‐displacement curves for each Intact, DP, and AF + EP injuries for (B) 4 month, (C) 12 month, and (D) 24 month motion segments. Significant differences were detected for axial properties, including (E) Compressive Stiffness, (F) Tensile Stiffness, (G) Range of Motion (ROM), and (H) Hysteresis. Statistical analysis compared differences across ages, injury types, and interactions (Age × Injury) using two‐way ANOVA, and if significant, then Tukey's post hoc testing was done. Significant differences are indicated as *p < 0.05, **p < 0.01, ***p < 0.0001 and ****p < 0.00001.

3.2. Injury Decreased Time Constants in 4 and 12 Month IVDs and Increased Displacement Loss in 24 Month IVDs

The creep test was conducted to evaluate the viscoelastic properties of the IVD, specifically assessing the fast and slow time constants, as well as displacement loss (Figure 3A). Representative graphs show Intact IVDs (gray), DP injuries (red), and AF + EP injuries (blue) across 4 months (Figure 3B), 12 months (Figure 3C), and 24 months (Figure 3D) cohorts. Age and injury increased displacement loss from elastic compression, with 24 months IVDs showing a significantly higher loss in DP (vs. Intact) and AF + EP (vs. Intact) groups (Figure 3E). Age did not affect fast creep displacement loss, but injury increased displacement loss (Figure 3F), especially in the 24 months DP injury (vs. Intact) group. Injury significantly reduced the fast time constant in both DP and AF + EP groups, particularly at 12 months with a significant interaction between age and injury (Figure 3G). Although age and injury individually did not impact the slow creep displacement loss, there was a significant age‐injury interaction. Injury had no effect on 4 months IVDs but led to a decrease in 12 months IVDs and an increase in 24 months IVDs (Figure 3H). Age exhibited a trend (p = 0.07) toward increasing the slow time constant, whereas injury significantly reduced it in 12 months IVDs for both the DP and AF + EP groups compared to Intact IVDs. Additionally, the slow time constant decreased with DP injury in 24 months IVDs compared to Intact IVDs (Figure 3I). Injury reduced the slow time constant allowing water to leave faster and the tissue to creep more quickly, and it also reduced the fast time constant allowing lateral bulging to occur more rapidly. A significant interaction between age and injury was observed in total displacement, with DP injury (vs. Intact) exacerbating the total displacement loss exclusively in 24 months IVDs (Figure 3J).

FIGURE 3.

Injury resulted in a greater displacement loss in 24 month IVDs than in 4 and 12 month IVDs. (A) Representative axial creep force‐time curve fits a five‐parameter exponential model. (B–D) Representative axial creep curves for each Intact, DP and AF + EP injuries for (B) 4 month, (C) 12 month, and (D) 24 month IVDs. Significant differences were detected for axial creep properties, including (E) Elastic Displacement, (F) Fast Creep Displacement, (G) Fast Time Constant, (H) Slow Creep Displacement, (I) Slow Time Constant, and (J) Total Displacement. Statistical analysis compared differences across ages, injury types, and interactions (Age × Injury) using two‐way ANOVA and if significant, then Tukey's post hoc testing was done. Significant differences are indicated as *p < 0.05, **p < 0.01, ***p < 0.0001 and ****p < 0.00001.

3.3. Injury Decreased Torque Range and Torsional Stiffness in 12 and 24 Month IVDs

Torsional properties were assessed from the 3rd cycle of the hysteresis curve, including torque range, torsional stiffness, NZ length, and hysteresis (Figure 4A). Representative graphs show Intact IVDs (gray), DP injuries (red), and AF + EP injuries (blue) across 4 months (Figure 4B), 12 months (Figure 4C), and 24 months (Figure 4D) cohorts. Although Intact IVDs across all ages exhibited a similar torque range, DP injury reduced the torque range by 86% in 12‐month IVDs and 88% in 24‐month IVDs, while having no effect in 4‐month IVDs (Figure 4E). DP injury reduced torsional stiffness by 68% in 12 months (p = 0.06) and 76% in 24‐month IVDs (Figure 4F). AF + EP injury (vs. Intact) increased NZ length, particularly in 24‐month IVDs (Figure 4G). Age reduced torsional hysteresis, with DP injury further decreasing it by 56% at 12 months and 76% (p = 0.06) at 24 months (Figure 4H). DP injury had a greater impact on torsional properties than AF + EP injury. Comparison between the 3rd and 19th cycles at all ages showed a significant reduction in torsional properties, suggesting accumulation of damage at later cycles, and therefore parameters were estimated using torsional loading cycle 3 (Figure 5). When evaluating torsional properties calculated from loading cycle 19, no significant effects of age or injury were detected, likely due to damage induced from the loading protocol which involved relatively large rotation angles (Figure S2).

FIGURE 4.

Injury caused a greater reduction in torque range and torsional stiffness in 12 and 24 month IVDs compared to 4 month IVDs. (A) Representative torque‐angular displacement curve showing the calculated torsion parameters. (B–D) Representative torque‐angular displacement curves for each Intact, DP and AF + EP injuries for (B) 4 month, (C) 12 month, and (D) 24 month IVDs. Torque properties, including (E) Torque Range, (F) Torsional Stiffness, (G) Torque NZ Length, and (H) Torsional Hysteresis. Statistical analysis compared differences across ages, injury types, and interactions (Age × Injury) using two‐way ANOVA, and if significant, then Tukey's post hoc testing was done. Significant differences are indicated as *p < 0.05, **p < 0.01, and ***p < 0.0001.

FIGURE 5.

Torsion properties were significantly affected by the torsional loading cycle, and this damage suggests early cycles are more suitable for analysis. Torque cycles were compared between the 3rd and 19th cycles for different torsional properties: (A) Torque range, (B) Hysteresis, and (C) Torsional stiffness in 4, 12, and 24 month IVDs. Statistical analysis to compare differences in 3rd and 19th cycle outcomes across all ages was performed using one‐way ANOVA. Significant differences are indicated as *p < 0.05, **p < 0.01, ***p < 0.0001 and ****p < 0.00001.

For axial, creep, and torsion biomechanical outcomes, we further examined biomechanical results stratified by sex in cases where the data suggested a potential bimodal response (Figures S3 and S4). Our analysis indicated that sex differences did not account for the observed separation, suggesting that sex‐related effects on these properties are minimal or negligible.

3.4. Age Led to Decrease in Proteoglycan Content

FTIR spectra of 4, 12, and 24 month L5‐6 IVDs were analyzed to assess various compositional properties, including proteoglycan content, bound water, and collagen cross‐linking (Figure 6A). Proteoglycan content/Amide I ratio was significantly decreased in 24 month IVDs with 59.4% and 56.2% compared to 4 and 12 month IVDs, respectively (Figure 6B). No significant changes were detected in bound water/Amide I ratio (Figure 6C), collagen cross‐linking (Figure 6D), Amide II/Amide I (Figure S5A), or collagen integrity as detected with CH₂/Amide II ratio (Figure S5B).

FIGURE 6.

Age led to a decrease in proteoglycan content in IVDs. (A) Representative FTIR spectra of 4, 12, and 24 month L5‐6 IVDs showing proteoglycan, Amide I, Amide II, and bound water peaks. Different biophysical properties were measured for 4, 12, and 24 month IVDs: (B) Proteoglycan/Amide I, (C) Bound water/Amide I, (D) Collagen Cross‐linking. Statistical comparisons across ages were performed using one‐way ANOVA. Significant differences are indicated as *p < 0.05, **p < 0.01, and ***p < 0.0001.

3.5. Age Led to Decrease in Proteoglycan and Increase in Collagen Denaturation

Safranin O staining was observed in the entire NP and endplate (EP) compartments of the lumbar IVD at 4 months of age (Figure 7A). A significant portion of inner AF (IAF) stained intensely with Safranin O, while the outer AF (OAF) lamellae, which are more collagenous, were rich in Fast Green staining in the 4‐month mouse lumbar IVD. In 4‐month IVDs, even the vertebral epiphysis (VE) and cartilage EP (CEP) were highly stained with the Safranin O staining (Figure 7A). At 12 months, the NP continued to stain intensely with Safranin O. Safranin O stained either the entire VE or slightly the innermost CEP layer adjacent to the NP at 12 months of age. Additionally, there was a formation of a center of ossification in the CEP. The AF continued to show two distinct regions with inner layers rich in Safranin O positive staining and outer layers rich in Fast Green staining (Figure 7A). The Safranin O staining appears more diffused in the 12‐month IVDs between the OAF and IAF, lacking a clear boundary between the two regions. At 24 months of age, the NP region of lumbar IVD lost the cell band and was still Safranin O positive but also showed Fast Green staining in IAF, indicating age‐related loss of proteoglycans and an increase in collagen content (Figure 7A). The EP region was devoid of any Safranin O staining and stained positive for Fast Green staining, indicating the loss of proteoglycans in the EP by 24 months of age in mouse lumbar IVDs. At 24 months of age, the AF of lumbar IVDs was majorly Fast Green positive, and very few inner AF layers showed mild to no Safranin O staining, indicating loss of proteoglycans in the AF by 24 months of age. Additionally, the CEP exhibited a significant reduction in Safranin O staining, indicating proteoglycan loss, and displayed a more collagenous nature, characterized by predominant Fast Green positivity and an increase in the center of ossification.

FIGURE 7.

Age led to a decrease in proteoglycan and an increase in collagen denaturation in IVDs. Representative images of mid‐coronal sections of mouse L5‐6 IVDs stained with (A) Safranin O and Fast Green, (B) Picrosirius Red followed by polarized microscopy captured at 10× and 40× magnification, (C) IVD height, and (D) IVD area for 4, 12 and 24 month IVDs. Statistical comparisons across ages were performed using one‐way ANOVA. Significant differences are indicated as *p < 0.05, **p < 0.01, and ***p < 0.0001. CEP‐Cartilaginous Endplate; IAF‐Inner Annulus Fibrosus; OAF‐Outer Annulus Fibrosus; NP‐Nucleus Pulposus; VE‐Vertebral Endplate.

Picrosirius Red staining followed by polarized microscopy was used to analyze the age‐related change in collagen content and collagen type in the mouse lumbar IVDs. Following Picrosirius Red staining, thick and Type I collagen fibers were reported to appear orange‐red, and thin or Type III collagen fibers were reported to appear green or yellowish green under polarized microscopy [49, 50, 51]. At about 4 months of age, the NP did not show collagen content under polarized microscopy (Figure 7B). Polarized microscopy identified thick and well‐aligned red collagen fibers in the entire AF at 4 months of age. The cartilaginous EP also showed collagen fibers at 4 months of age. However, the collagen fibers at the AF‐EP interface were a mix of red and greenish fibers, indicating ongoing maturation at this age. NP did not show collagen content even at 12 months of age (Figure 7B). AF continued to be majorly formed of thick collagen fibers, as noted by the intense red staining, but some thinner green collagen fibers were also noticed in the AF at 12 months of age. The cartilaginous EP was majorly red, indicating thicker and well‐formed collagen. By 24 months of age, the amount of collagen content in the entire IVD was observed as green, indicating thinner collagen and more Type III collagen compared to Type I observed in 4 and 12 month mouse IVDs (Figure 7B). The NP showed signs of thin collagen lamellae. The AF showed remarkable changes with age, as more green and thin fibers were observed in the inner and significant portion of the entire AF by 24 months. EP adjacent to AF did not detect much collagen at 24 months. Thinner and weaker collagen, indicated by a yellowish‐green signal, was detected in the CEP adjacent to NP at 24 months of age (Figure 7B).

3.6. Age Reduced IVD Height and Increased Cross‐Sectional IVD Area

IVD height significantly decreased in 24 months compared to 4 and 12 months IVDs by 13.6% and 11.6%, respectively (Figure 7C). The IVD area increased in 24 months IVDs compared to 4 months IVDs by 12% (Figure 7D), while no significant changes were detected in IVD circularity (Figure S6C) or convexity index (Figure S6D) with age. IVD height decreased with age, showed a trend toward significantly decreasing with spinal level (p = 0.05), and showed no significant interaction between age and level (Figure S6C). No change in IVD area with age and spinal level was observed (Figure S6D). At 4 months, L5‐6 circularity increased by 45% compared to L1‐2 and 21% compared to L3‐4. At 12 and 24 months, L5‐6 circularity increased by 27% and 22% compared to L1‐2 (Figure S6E). No significant changes in IVD convexity index were observed with age or spinal level (Figure S6F).

4. Discussion

The high prevalence of age‐related back pain motivated this study into the interaction between age and injury on the ex vivo biomechanical properties of mouse lumbar IVD motion segments with 2 distinct injury phenotypes. The most important finding is that injury causes more severe changes to mouse lumbar IVD biomechanical properties in aged mice compared to younger mice. The largest changes were in compressive stiffness, elastic displacement, and creep time constants, suggesting injuries that impact NP pressurization have a larger effect in 12 and 24‐month‐old IVDs. Torsional properties also were affected more by injuries in 12 and 24‐month IVDs. Together, these results highlight the interaction between age and injury with the largest effects on biomechanical dysfunction in 12 and 24‐month mice. Importantly, injury type also significantly affected mouse spine biomechanical properties, with DP exhibiting greater changes than AF + EP, demonstrating that differing IVDD phenotypes cause distinct biomechanical property changes.

The effects of the injury on ex vivo biomechanical properties were most severe in 24‐month IVDs and often undetectable in 4‐month IVDs. Older IVDs have less GAG and more denatured collagen than young IVDs [19], as well as increased elastic displacement loss. Since we have found that 4‐month IVDs have higher amounts of proteoglycans than 24‐month IVDs, we believe that the high proteoglycan content along with well‐aligned stronger collagen of AF allowed the 4‐month IVDs to resist the forces and swell to ‘re‐tension’ the AF fibers. Based on measured dimensions of the IVDs and those of the 26G needle (outer diameter = 0.46 mm and inner diameter = 0.26 mm), we estimate that approximately 20% of the NP may have been extracted by the bore of the needle during the DP injury procedure. The AF + EP injury procedure likely caused some removal of NP, although it is believed to be less than the DP. Cyclic axial loading may have caused further NP herniation in the 4‐month DP group. It is possible that the 26G needle puncture caused the highly gelatinous NP of the 4‐month IVDs to herniate more fully than the more fibrous NP of the older IVDs, so that the measured properties of the injured 4‐month IVDs were more representative of the stiffer AF structure with relatively few biomechanical changes from the intact condition. Future comprehensive studies are required to examine structural, biochemical, and biomechanical changes in each compartment of the IVD post‐injury to provide better insights into the role of each compartment in observed biomechanical changes with injury and age. Therefore, this higher proteoglycan content and stronger collagen of young IVDs limit the changes in compressive stiffness and torque range from injury. Nevertheless, injury significantly decreased the slow time constant in 4‐month IVDs, showing injury still shifted load‐carrying mechanisms with more rapid creep. The reduction in the slow time constant is similar to the increase in hydraulic permeability observed in puncture injury of rat tail IVDs evaluated with stress‐relaxation tests ex vivo using a poroelastic model [52, 53, 54]. Previous work also found that injury caused biomechanical changes which were similar to coccygeal IVDs in 6.5–9‐month mice [30]. Histological analyses following injury, which are planned for a future in vivo study, are required to examine structural changes in both the NP and AF post‐injury in order to provide more insights into the observed biomechanical changes with injury and age.

Interestingly, age significantly reduced tensile stiffness [26], which is consistent with results from the 2‐way ANOVA. Furthermore, injury‐induced IVDD was demonstrated to be more severe than age‐related IVDD in a rat model [55]. Taken together, the studies confirmed known age‐related and increased severity of injury‐induced changes to IVD biomechanical properties, and we further demonstrated that these factors interact with injury, causing more severe biomechanical dysfunction in 12 and 24‐month mouse IVDs. However, an opposite trend was observed for DP injury, with reduced tensile stiffness in 4‐month IVDs but not in 12 and 24‐month IVDs. Since tensile stiffness was the only parameter to show greater sensitivity to injury in young IVDs, it is believed that a different mechanism is at play. Specifically, the tensile modulus of the intact 4‐month animals is larger than all other groups. We believe this difference in tensile modulus, potentially caused by collagen denaturation, is low at skeletal maturity and accumulates with aging. However, AF collagen and crosslinking changes with aging may also be additional factors, as is shown by the Picrosirius Red staining in our study [19, 31, 56]. Therefore, the DP injury, which affects the AF more severely than the AF + EP injury, had the largest impact on the youngest IVDs with the largest tensile modulus.

Clinical studies indicate that geriatric patients treated for lumbar IVD herniation tend to experience worse postoperative outcomes and persistent back pain after surgery compared to younger adults [57]. Aging is also a significant contributor to the progression of spondylolisthesis, a condition closely linked to spinal instability [58, 59]. While spondylolisthesis is influenced by multiple factors, including IVDD, facet joint orientation, and facet arthritis, research indicates that IVDD plays a primary role in initiating the condition [60]. Our results show biomechanical and biochemical changes with age to provide a potential mechanism of action for the inferior clinical outcomes in older people. We found age decreased IVD height and increased IVD area, likely due to reduced GAG content shown by loss of Safranin O staining leading to loss of pressurization and a flatter, expanded IVD area, consistent with aged human IVDs [61, 62, 63]. We also report age accumulated thinner, denatured collagen fibers, further compromising IVD biomechanical function. Lastly, injury caused more severe biomechanical property changes in aged IVDs with all significant findings pointing to greater IVD deformations under load. Greater IVD deformations under the load predispose to conditions such as IVD height loss, increased facet loading, and spondylolisthesis, which result in less space for surrounding nerves and are causes of pain and disability. For example, torsional properties in 12 and 24‐month IVDs were most strongly affected, and this is consistent with concordant pain at discography correlating with increased axial rotation angles [64]. IVD healing capacity also diminishes with age so that IVD injury in older patients can more rapidly lead to IVDD and related pathologies [31, 65, 66]. As such, our study, in conjunction with literature, supports the concept that older spines experience greater biomechanical dysfunction compared to younger spines following injury and shows that one mechanism for these observations is greater deformation under axial or rotational loads.

This study compared different injury conditions. The DP injury involved full‐width midline puncture entering the anterior AF, traversing through the NP, and exiting through the posterior AF to cause extensive disruption to the AF and NP compartments to simulate severe IVDD. In contrast, the AF + EP injury affected only one side of the AF with a more mild disruption to the EP and NP tissues, to simulate IVD herniation with EP junction failure [38], which was a more localized injury and milder IVDD condition. Therefore, DP exhibited more severe changes in biomechanical properties than AF + EP since it was a more severe injury. Both injuries were similarly affected by age. Damage accumulates in IVDs, and unrepaired AF defects can progress to more severe IVDD [9, 67]. This highlights the compounded challenges faced by older patients with advanced IVDD, where both age‐related compositional changes and the severity of IVD injury exacerbate biomechanical dysfunction and further diminish healing potential. Research has demonstrated distinct patterns of biomechanical dysfunction due to differing IVD injuries and IVDD phenotypes, to underscore the need for tailored treatments based on the injury type to restore biomechanical function. This study augments this understanding to clarify that biomechanical dysfunction is directly impacted by age, highlighting the clinical need to tailor treatment approaches for elderly patients with lumbar spine disorders and the critical research need for improved IVD repair strategies, especially for older patients.

Interpretations of this ex vivo biomechanical study are limited to acute injury responses, as there is no healing or inflammatory response in this model, and an in vivo study is required to investigate the healing and pain‐related responses from these injuries and how they are affected by age. A microscope was used for injury creation to minimize variability and create two distinct phenotypes relevant to human IVDD and herniation. However, some variability is inevitable due to the small size of the mouse lumbar IVDs and the fact that human clinical conditions typically develop gradually over the years, rather than from a single acute injury. Stratification of biomechanical results by sex did not suggest that data separated by sex; therefore, the results are presented combined by sex, and this study was not designed to detect small effects of sex should they be present. We did not detect any effects of spinal level on IVD area and only minor effects for IVD height, we systematically rotated the placement of Intact, AF + EP, and DP across levels to adjust for potential level‐specific effects since the same size needle was used for all levels. We identified several biomechanical property changes that were identified with age and injury and with their interaction, yet the small size of mouse lumbar IVDs results in some variability that could limit the capacity to detect more subtle effects than those detected in this study.

Histochemical staining of histological sections of the lumbar IVDs from the 4, 12, and 24 month mice provides further insights into the region‐specific age‐related biochemical changes observed by biochemical testing. Safranin O and fast green staining validated age‐related decline in the proteoglycan content observed by FTIR in each compartment of the IVD, with the earliest changes observed in the EP and AF regions by 12 months of age. Furthermore, by 24 months of age, even the NP region had markedly reduced proteoglycan content. However, the decline in proteoglycan was associated with poor quality of collagen in all compartments of the IVD, supporting the poor mechanical properties and dramatic response to injury observed in the IVDs from 12 and 24 month mice. Moreover, with the age‐related decline in proteoglycan content and collagen denaturation, a decline in the IVD height and increase in IVD area were also observed.

In summary, we found mouse lumbar spine ex vivo biomechanical properties were affected by age and injury, with injury having a greater effect. Furthermore, injury interacted with age to induce more severe biomechanical changes in middle‐aged and geriatric IVDs, likely because of the proteoglycan loss and collagen denaturation measured with age. Further, DP caused greater changes than AF + EP, indicating that structural disruption affecting AF and NP compartments mimicking severe IVDD caused greater changes than AF + EP injuries mimicking herniation with EP junction failure. We conclude that lumbar spinal injury interacts with age in impacting the acute ex vivo biomechanical properties and points to greater deformations under load in aged and geriatric spines to provide a mechanism for inferior healing, pain, and disability responses.

Author Contributions

Conceptualization: N.B. and J.C.I. Study design: N.B. and J.C.I. Resource acquisition: N.B., J.T., C.L.D., and J.C.I. Performed experiments: N.B., J.T., A.B., B.W., P.N., A.P.H., and V.S. Analyzed data and prepared figures: N.B., J.T., A.B., C.J., B.W., A.D., C.L.D., and J.C.I. Interpreted results and contributed to figure edits: N.B., A.B., J.T., C.J., B.W., A.D., P.N., A.P.H., V.S., J.R.W., D.V., C.L.D., A.C.H., and J.C.I. Funding acquisition: J.C.I. Project administration: N.B. and J.C.I. Supervision: N.B. and J.C.I. Writing – original draft: N.B., J.T., B.W., C.L.D., and J.C.I. Writing – review, editing, and approval: N.B., A.B., J.T., C.J., B.W., A.D., P.N., A.P.H., V.S., J.R.W., D.V., C.L.D., Y.L., A.C.H., and J.C.I.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Injury resulted in an increased NZ length and decreased NZ stiffness across all age groups. Axial cyclic properties in mouse lumbar IVD, (A) Neutral Zone (NZ) Length and (B) NZ Stiffness. Statistical analysis compared differences across ages, injury types, and interactions (Age × Injury) using two‐way ANOVA and if significant then Tukey's post hoc testing was done. Significant differences are indicated as *p < 0.05, **p < 0.01, and ***p < 0.0001. Post hoc comparisons did not detect statistically significant differences (p > 0.05).

Figure S2: Age and injury did not show significant changes in the torsional properties of the IVD during the later 19th loading cycle. Torsional properties, (A) Torque Range, (B) Torque Stiffness, (C) Torque NZ Length and (D) Torsional Hysteresis. Statistical analysis compared differences across ages, injury types, and interactions (Age × Injury) using two‐way ANOVA and if significant then Tukey's post hoc testing was done. Significant differences are indicated as *p < 0.05, **p < 0.01, and ***p < 0.0001. Post hoc comparisons did not detect statistically significant differences (p > 0.05).

Figure S3: Sex had little to no impact on the axial compression and tension biomechanical properties of IVDs. Axial properties: Compressive Stiffness: (A‐Male+Female), (A′‐Male), (A″‐Female). Tensile Stiffness: (B‐Male+Female), (B′‐Male), (B″‐Female). Range of Motion: (C‐Male+Female), (C′‐Male), (C″‐Female). Hysteresis: (D‐Male+Female), (D′‐Male), (D″‐Female). Statistical analysis compared differences across ages, injury types, and interactions (Age × Injury) using two‐way ANOVA and if significant then Tukey's post hoc testing was done. Significant differences are indicated as *p < 0.05, **p < 0.01, and ***p < 0.0001. Post hoc comparisons did not detect statistically significant differences (p > 0.05).

Figure S4: Sex had little to no impact on the creep and torsional biomechanical properties of IVDs. Creep properties: Fast Creep Displacement: (A‐Male+Female), (A′‐Male), (A″‐Female). Total Displacement: (B‐Male+Female), (B′‐Male), (B″‐Female). Torsional Stiffness: (C‐Male+Female), (C′‐Male), (C″‐Female). Statistical analysis compared differences across ages, injury types, and interactions (Age × Injury) using two‐way ANOVA and if significant then Tukey's post hoc testing was done. Significant differences are indicated as *p < 0.05, **p < 0.01, and ***p < 0.0001. Post hoc comparisons did not detect statistically significant differences (p > 0.05).

Figure S5: Age did not cause any significant changes in collagen type and collagen integrity. Biophysical properties measured using FTIR: (A) Amide II/Amide I (collagen type) and (B) CH₂/Amide II (Collagen Integrity). Statistical comparisons across ages were performed using one‐way ANOVA. Significant differences are indicated as *p < 0.05, **p < 0.01, and ***p < 0.0001.

Figure S6: Age did not cause significant changes in IVD circularity and convexity index. Age effect: (A) IVD Circularity, and (B) IVD Convexity Index in 4, 12, and 24 month IVDs. Age and level interaction: (C) IVD Height, (D) IVD Area, (E) IVD Circularity, and (F) IVD Convexity Index. Statistical comparisons were performed using (A, B) one‐way ANOVA across ages and (C–F) two‐way ANOVA across ages, level and interactions (Age × level) and if significant then Tukey's post hoc test was done. Significant differences are indicated as *p < 0.05, **p < 0.01, and ***p < 0.0001.

Bhadouria N., Tiao J., Baburova A., et al., “Injury Induces More Severe Biomechanical Changes in Middle‐Aged and Geriatric Lumbar Spines in a Mouse Ex Vivo Model,” JOR Spine 8, no. 4 (2025): e70127, 10.1002/jsp2.70127.

Data Availability Statement

All data are available in the main text or the Supporting Information.