Low-Intensity Vibrations Partially Maintain Intervertebral Disc Mechanics and Spinal Muscle Area During Deconditioning

INTRODUCTION

The intervertebral disc (IVD) receives a variety of complex loads during daily activities, engendered from gravity and back musculature, that are diverse in magnitude, frequency and duration [1]. These loads are supported by the central nucleus pulposus which transfers axial load to the inferior endplate-vertebrae and distributes radial load to the outer collagenous annulus fibrosus. Prolonged, reduced dynamic loading, as occurs with immobilization, bed rest or exposure to microgravity, alters spine physiology [2–4] and remodels the intervertebral disc [5]. Ultimately, these changes unbalance the load distribution of the disc, potentially stimulating nociceptive nerves [6] and increasing the risk of mechanical IVD failure [7].

Although overnight bed rest reduces the spinal range of motion and swells the IVD [8], these alterations return to normal following brief reloading [9] and do not produce mechanical features associated with degeneration [10]. On the other hand, prolonged bed rest induces discal swelling and also differentially changes spinal musculature [11] such that 5 months of reambulation does not restore the discal morphology induced by 3 wk of bed rest [12]. Two months of bed rest increases the axial compressibility of the spine [3], which may increasingly strain the annulus as shown in human degenerated discs [13]. During bed rest, the posterior region incurs greater morphological changes [11] which may amplify functional alterations in this thinner and weaker region.

Animal models of degeneration offer unique opportunities to test the mechanical outcome of potential countermeasures to disc degeneration via invasive procedures which are not possible to perform in humans. Hindlimb unloading is a widely utilized animal model for simulating musculoskeletal, orthostatic and cardiovascular changes that arise from diminished functional weightbearing. Although hindlimb unloading primarily reduces axial spine loads, rather than removing most components of the complex mechanical environment, many outcome variables show similar IVD effects to those observed during spaceflight and degeneration, including decreased hydrophilic glycosaminoglycan content and discal hydrostatic pressure, altered collagen organization, and up-regulation of catabolic molecules [4,14,15]. Hindlimb unloaded discs also contain larger collagen fiber diameters in the posterior region [4] which, while priming the structure for tensile loading, may render the IVD less capable of recovering creep [16], and more susceptible to mechanical damage. Exposure to space flight for five days increases the stiffness of lumbar discs [17] while hindlimb unloading diminishes the speed of trunk rotation by reducing the electromyogram activity of the internal oblique, increasing the number of slow type I fibers, and reducing the metabolic activity of fast type II fibers [18]. However, the effects of spaceflight or hindlimb unloading on the mechanics of the rodent IVD, back musculature, and how they relate are largely unknown.

Effective countermeasures following disc deconditioning have not been identified. Mechanical signals applied as whole body vibrations may attenuate changes in IVD morphology, lower the incidence of back pain, and prevent muscle changes during spinal unloading [19,20]. In rats standing on a vertically vibrating plate in erect posture, facilitating the transmission of the oscillatory signal from the feet into the spine, 90 Hz low-level vibrations maintained glycosaminoglycan content and height, but not area, of the IVD compromised by hindlimb unloading [21]. While this study demonstrated the benefits of 90 Hz erect body vibrations to the composition and height in rats during unloading, it was not clear whether the mechanical behavior of the disc also benefited from the vibrational countermeasure during unloading. In an effort to address this question, here, we evaluated the consequences of 90 Hz low-level vibrations on disc mechanics and its relation to disc morphology following immobilization. Further, we assessed spinal muscles to clarify the role of endogenous loading in modulating the potential benefits of low-intensity vibrations.

MATERIALS AND METHODS

Experimental Design

This study was reviewed and approved by the Institutional Use and Care Committee of Stony Brook University. Similar to a previous study [21], female Sprague-Dawley rats (18 wk of age) were assigned to the following groups upon two weeks of acclimatization in our animal facility. One group of rats was hindlimb unloaded for 4 wk without interruption (HU, n=8). A sham control group of rats (HU+SC, n=8) was also subjected to HU but unloading was interrupted by brief periods (15 min/day) of upright posture, facilitated by standing in a vertical plastic cylinder (height: 30.5 cm, inner diameter: 10.2 cm). A third group of experimental rats (HU+90, n=8) was also hindlimb unloaded and in these animals, unloading was interrupted by standing in the upright cylinder on a plate that was vertically oscillating (15 min/day, peak acceleration = 0.2 g, frequency = 90 Hz). With a surface area of ~1000 cm², we were able to fit six rats in their cylinders on the vibrating plate. The plate has been used and characterized previously [20–22]. Normally ambulatory age-matched (AC) and baseline (BC) rats were used as controls (n=8, each). Animals had access to standard rat chow and water ad libitum.

Micro Computed Tomography of IVD and Muscle

An in vivo two-dimensional x-ray image at 0-and 4-wk was taken of the spine (“scout view” of VivaCT 75; Scanco, Brüttisellen, Switzerland) to determine the change in anterior, center, posterior and average IVD height. Average IVD height was calculated as the mean of anterior, center and posterior height. Following each scout view image, the inferior L5 vertebra to the superior L4 vertebra region was imaged at a resolution of 78 μm (45 kV, 177 μA, 300 ms integration time). An automated algorithm [23] segmented muscle from bone to determine muscle volume including the paraspinal (e.g., erector spinae, multifidus, quadratus lumborum, which could not be individually distinguished due to low in-plane contrast) and psoas. Since both unloading and vibrations coupled with resistive exercise differentially alter the spinal musculature [19], the psoas and paraspinal muscles were contoured in 3 consecutive images at inferior L4 and L5. After sacrifice, the lumbar spine was stored en bloc at −20 °C.

Mechanical Testing

Prior to mechanical testing, the L4-L5 bone-disc-bone segment was excised, facets and extraneous tissue removed, and hydrated in 1× PBS for 18 h at 4 °C. The L4 vertebra was gripped by custom-made microvises and the L5 vertebra was fixed with bone cement into a cylinder. Once secured, the sample was immersed in 1× PBS. Controlled mechanical tests were performed on the L4-L5 segment to describe physiological loading modalities (i.e., axial compression-tension, creep and torsion) [24–27]. An Instron 5542 testing system (Instron, Canton, MA) applied a load range of -6 N to 3 N at a frequency of 0.1 Hz. These loads were selected to be physiological and to be in the linear-region of the force and displacement response. After twenty compression-tension cycles, a 1 sec ramp from 0 to 6 N compression was applied and maintained for 1 h. Immediately following creep testing and while maintaining the 6 N compression load, a stepper-motor (AM15E0045; Faulhaber, Clearwater, FL) superimposed 10 cycles of torsion to ±8 ° rotation at 1.6 °/s [27].

Axial IVD Geometry

Following mechanical testing, the L4-L5 IVD was bisected and imaged using a stereomicroscope (Leica MZ6, Bannockburn, IL). The horizontal image was analyzed using a custom Matlab algorithm [28] to determine the antero-posterior width for the entire disc (W_AP) and nucleus pulposus (N_AP), the lateral width for the entire disc (W_L) and nucleus pulposus (N_L), as well as the area for the annulus fibrosus and nucleus pulposus. Assuming the nucleus pulposus supports negligible torsion, polar moment of inertia (J) of the IVD was calculated as $J=(\pi ∕64)\left[W_L^3{W}_{AP}+W_{AP}^3+W_L-\left(N_L^3{N}_{AP}+N_{AP}^3{N}_L\right)\right]$ [25, 27]

Mechanical Data Analysis

Mechanical data were analyzed by custom Matlab routines. The trilinear fit model determined the compressive, tensile and neutral zone stiffness of the L4-L5 motion segment during the 20^th compression-tension cycle [26]. Briefly, the compressive and tensile loading curves were isolated and a 5^th order polynomial was fit. The lowest derivative of the curve represented the neutral zone stiffness and the maximum derivative measured between 80% and 100% of the maximum load in the compressive and tensile direction constituted the compressive and tensile stiffness, respectively. Compressive range of motion (ROM) was normalized to IVD height. Similarly, the preconditioned 10^th cycle of torsion was fit with a trilinear model to obtain neutral zone (K_NZ) and linear region (K) stiffnesses in both the clockwise and counterclockwise. The apparent torsional neutral zone modulus (G_NZ) and linear region modulus (G) were calculated as G_i=K_ih/J (MPa/°) where h was the IVD height and J was the polar moment of inertia. K and G represented the mean from the linear regions of the clockwise and counterclockwise rotations. After the 1 sec step load of 6 N of creep force (F_C), 1 h of creep displacement was fit to a rheological model [29]: d(t)/F_C=1/S₁(1−e^{−S₁t/n₁}+1/S₂(1−e^{−S₂t/n₂}+1/S₃ where S_1,2 are the elastic damping coefficients, S₃ is the instantaneous elastic stiffness, η_1,2 are the viscous damping coefficients, and τ_1,2 are the time constants. The apparent elastic and viscoelastic moduli (MPa) were determined by normalizing the stiffness and damping coefficient by IVD geometry (height/area).

Statistics

Paired t-tests compared temporal changes in disc height and muscle morphology within each group. Unpaired t-tests compared age-related changes between BC and AC rats. The rationale of including a sham control group (HU+SC) was to separate the effect of weight bearing from the effect of vibrations during weight bearing. As expected, t-tests confirmed demonstrated that no outcome variable (except for torsional stiffness and modulus) was significantly different between HU and HU+SC groups (i.e., no effect of weight bearing). In an effort to preserve statistical power and to avoid raising the probability of committing a type II error (by including a group in the comparisons that was not critical for addressing the hypotheses), HU and HU+SC were pooled and referred to as HU±SC. The pooling was not performed for analyses involving torsional stiffness and modulus. One-way ANOVAs with post-hoc Tukey tests compared the temporal change in IVD height and muscle morphology, and endpoint mechanical properties of the IVD between groups.

For each group of rats, linear regressions were performed between the neutral zone modulus and the change in posterior IVD height. Morphology of the posterior region was selected because it is more susceptible to mechanical deterioration than the anterior region and its altered mechanics relate well to degeneration [13]. Linear regressions also tested for associations between muscle area and mechanical variables, using each group mean as a data point. Due to high variability in the mechanical outcomes, one-sample t-tests with a significance threshold of 0.01 were used to remove outliers from each group [30]. This test was applied to all groups without bias. If there were two outliers in a group that satisfied the above rule, only the farthest outlier was removed. Data were expressed as mean ± standard deviation (mean ± SD). Relative differences between sample means were denoted as percent difference ± SD of the sampling distribution of the relative difference (relative standard error of the difference). Statistical significance was considered at p < 0.05 (SPSS 18; SPSS Inc., Chicago, IL), unless otherwise stated. One author (SJ) owns (provisional) patents regarding the application of vibrations to the musculoskeletal system.

RESULTS

IVD Geometry

Discs of rats that were hindlimb unloaded with and without brief upright weight bearing were hypotrophied. Compared to AC, change in IVD height of the HU±SC rats was smaller by 96±24% (p=0.005) (Figure 1). Compared to AC, area of the nucleus pulposus of HU±SC discs was less by 39±8% (p<0.001) (Figure 2). High-frequency vibrations prevented changes to IVD morphology. Change in mean height of HU+90 IVD was greater than of HU±SC by 2052±893% (p=0.016) (Figure 1) with no difference in baseline height of the IVD of HU±SC (0.83±0.07 mm) and HU+90 (0.82±0.07 mm). Compared to HU±SC, nucleus pulposus area of HU+90 discs was greater by 35±12% (p=0.016) (Figure 2). There were no differences between the morphology of IVD from AC and HU+90 rats (p≥0.108). Annulus fibrosus area was not different between any groups (p=0.949).

Figure 1.

Change of intervertebral disc (IVD) height from baseline of age-matched (AC, n=7), hindlimb unloaded, with or without brief weightbearing (HU±SC, n=14) and hindlimb unloaded with brief weightbearing and 90 Hz oscillations (HU+90, n=7); mean+SD. The introduction of 90 Hz mechanical signals to upright posture maintained the normal change in IVD height. †: significant difference vs. AC, ‡: significant difference vs. HU±SC; p <0.05

Figure 2.

Nucleus pulposus area of age-matched (AC, n=7), hindlimb unloaded with or without brief weightbearing (HU±SC, n=14), and hindlimb unloaded, with brief weightbearing and 90 Hz oscillations (HU+90, n=7) IVD; mean+SD. Compared to AC, IVD of hindlimb unloaded rats, with or without upright loading, had a smaller nucleus pulposus area. The application of 90 Hz vibrations maintained normal nucleus pulposus area. †: significant difference vs. AC, ‡: significant difference vs. HU+SC; p <0.05

Compression-Tension

Hindlimb unloading altered the axial compression properties of the L4-L5 motion segment. Compared to AC, compressive modulus of HU±SC IVD was greater by 26±8% (p=0.020) and axial NZ modulus of HU±SC discs was smaller by 60±23% (p=0.043) (Figure 3). A similar increase was noted with the axial NZ stiffness of HU±SC discs but not with the compressive stiffness (Table 1). Tensile modulus and stiffness was not different between the 4 wk groups (p≥0.432).

Figure 3.

Compressive, tensile and neutral zone modulus of the L4–L5 segment of age-matched (AC, n=7), hindlimb unloaded with or without brief weightbearing (HU±SC, n=14) and hindlimb unloaded with, brief weightbearing and 90 Hz oscillations (HU+90, n=7) rats; mean+SD. Compared to AC, HU±SC segments had weaker neutral zone moduli and greater compressive moduli. HU+90 segments had greater neutral zone moduli and smaller compressive moduli than HU±SC segments. †: significant difference vs. AC, ‡: significant difference vs. HU±SC; p <0.05

Table 1.

Structural mechanical properties of the L4-L5 motion segment of BC, AC, HU±SC and HU+90 rats

Group	SCOM [N/mm]	STEN [N/mm]	SNZ [N/mm]	Com ROM [mm]	S1 [N/mm]	S2 [N/mm]	S3 [N/mm]	KNZ [N/mm°]	K [N/mm°]
BC	97±5	41±5	4.6±2.6	0.31±0.26	34±5	17±2	466±40	0.30±0.11	1.1±.4
AC	89±11	40±3	6.8±3.6	0.21±0.06	39±4	18±1	447±41	0.31±0.05	1.6±.3*
HU±SC	99±11	37±5	2.3±1.8†	0.28±0.11	36±7	18±1	534±59†	0.23±0.03†	NA
HU+90	82±13‡	40±5	6.714.3‡	0.28±0.13	38±6	17±2	447±50‡	0.20±0.05†	1.4±.2

Data are presented as mean±SD; n=7 for baseline (BC), age-matched (AC) and hindlimb unloaded 90 Hz oscillations (HU+90); n=14 for hindlimb unloaded with or without brief weightbearing (HU±SC).

NA – Not pooled due to significant difference between HU (1.3±0.3) and HU+SC (1.7±0.2)

^*AC vs. BC

^†significant difference vs. AC

^‡significant difference vs. HU±SC; p<0.05

The introduction of 90 Hz vibrations quelled the increase in L4-L5 IVD stiffness seen in unloaded rats. Compared to HU±SC, the compressive modulus of HU+90 IVD was smaller by 18±9% (p=0.047) and the NZ modulus of HU+90 IVD was greater by 185±74% (p=0.012) (Figure 3). The intervention provided a similar protection for the respective structural mechanical properties (Table 1) but no differences in elastic properties existed between the IVD of AC and HU+90 rats (p≥0.482).

Creep

Compared to AC, the long time constant τ₂ of HU±SC IVD was 101±32s longer (p=0.009) and the instantaneous elastic modulus S₃ was 25±8% greater (p=0.008) (Table 2). The structural equivalent of the instantaneous elastic modulus, stiffness S₃, of HU±SC discs was similarly greater than of AC (Table 1). Superposition of 90 Hz low-intensity vibrations provided few benefits to creep property changes induced by hindlimb unloading. Vibrations did not mitigate the greater τ₂ of HU±SC discs (p=0.858) which was not different from the τ₂ of AC (p=0.067) (Table 2). Contrarily, stiffness S₃ of HU+90 discs was smaller by 16±5% (p=0.004) (Table 1) but the modulus was not different to that of HU±SC (p=0.278) or AC (p=0.305) (Table 2). Compressive range of motion, short time constant τ₁, S₁, and S₂ were not different between the 4 wk groups (p≥0.303).

Table 2.

Material mechanical properties and time constants of the L4–L5 motion segment in BC, AC, HU±SC and HU+90 rats

Group	Com ROM [mm/mm]	S1 [MPa]	S2 [MPa]	S3 [MPa]	τ1 [s]	τ2 [s]	GNZ [kPa/°]	G [kPa/°]
BC	0.27±0.19	3.1±0.6	1.7±0.4	41.8±5.1	16.3±2.4	792±74	19.1±6.3	80±22
AC	0.24±0.04	3.3±0.5	1.6±0.2	38.4±5.7	19.8±2.5*	768±76	18.5±3.0	105±26
HU±SC	0.36±0.17	3.6±0.9	1.6±0.2	48.0±7.2†	18.9±4.2	900±55†	15.9±3.0	NA
HU+90	0.31±0.15	3.5±1.0	1.7±0.2	43.4±4.5	19.5±4.4	884±8l†	14.2±3.5	100±12

Values are presented as mean±SD. Sample size for HU±SC was n=14 and n=7 for all other groups was

Note η=S_i. τ_i

NA – Not pooled due to significant t-test; values for HU and HU+SC were 97±13 and 112±24.

^*significant difference between AC vs. BC

^†significant difference vs. AC, p<0.05

Torsion

Torsional neutral zone stiffness K_NZ of AC discs was smaller than of HU±SC by 25±6% (p<0.001) but torsional stiffness K (p≥0.067) (Table 1), torsional NZ modulus G_NZ (p≥0.059) (Table 2), and modulus G (p≥0.105) of the L4-L5 IVD were not different between groups. Vibrations did not maintain K_NZ during unloading. There was no difference in K_NZ between HU+90 and HU±SC (p=0.161) but was 37±8% smaller (p<0.001) in HU+90 than AC (Table 1).

Spinal Muscle Morphology

Contrary to the lack of longitudinal change in AC (Figure 4, Table 1), HU±SC rats gained 17±15% (p<0.001) and 12±20% (p=0.038) psoas muscle area at L4 and L5, and lost 14±9% (p<0.001) paraspinal area at L4. Compared to AC, HU±SC rats lost three times (p<0.001) more paraspinal muscle area at L4 and no difference in change at L5 (p=0.220). Change in area of the psoas muscle at L4 and L5 was not different among the 4 week groups. Vibrations at 90 Hz superimposed on brief upright posture provided partial benefits to the altered spinal muscle size during hindlimb unloading. HU+90 animals did not gain psoas muscle area at either vertebral level (p≥0.138) (Figure 4). However, HU+90 rats lost 17±2% and 16±2% paraspinal muscle area at L4 and L5 (p<0.001) and, compared to HU±SC, HU+90 rats lost 74±38% more paraspinal muscle area at L5 (p=0.029). Compared to AC, the change in paraspinal muscle area of HU+90 rats was smaller by 425±71% and 591±87% at L4 and L5 (p<0.002). Over 4 wk, unloading groups gained lumbar muscle volume (p≤0.002) (Table 3).

Figure 4.

Change in paraspinal muscle area at L4 and L5 of age-matched (AC, n=7), hindlimb unloaded with or without brief weightbearing (HU±SC, n=14) and hindlimb unloaded with brief weightbearing and 90 Hz oscillations (HU+90, n=7) animals; mean+SD. Unloading with and without upright loading increased the psoas area at L4 and L5. At L4, paraspinal area of HU±SC was less than in AC rats. At L5, paraspinal area of HU+90 was less than in HU±SC and AC rats. *: significant difference vs. baseline of same animal, †: significant difference vs. AC, ‡: significant difference vs. HU±SC; p <0.05

Table 3.

Muscle volume between L4 and L5 of AC, HU±SC and HU+90 rats over 4 weeks [cm³]

Time	AC	HU±SC	HU+90
0 week	91.4±2.3	93.4±2.8	92.7±3.2
4 week	94.5±2.4*	95.7±3.3*	95.7±2.9*
Relative Change [%]	3.7 ± 2.1	1.9 ± 1.1	3.2 ± 2.1

Data are presented as mean±SD; n=7 for age-matched (AC) and hindlimb unloaded with brief 90 Hz oscillations (HU+90); n=14 for hindlimb unloaded with or without brief weightbearing (HU±SC).

^*significant difference vs 0 week, p<0.05

Relationship between Disc Mechanics and Geometry or Musculature

There was a moderate negative relationship between axial neutral zone modulus and change in posterior disc height in the AC group (R²=0.60, p=0.040) (Figure 5). Hindlimb unloading with and without brief weightbearing uncoupled the relationship between disc mechanics and changing geometry (R²=0.00, p=0.905). The application of 90 Hz vibrations normalized the relationship between axial neutral zone modulus and change in posterior disc height (R²=0.65, p=0.028). Across all groups, mean paraspinal muscle area was positively and highly related (R²=0.99, p<0.001) to torsional neutral zone stiffness of the L4-L5 IVD. No other significant relationships were detected between mean muscle area and material properties of the L4-L5 IVD.

Figure 5.

Regressions between neutral zone modulus and change in posterior disc height of age-matched (AC, n=7), hindlimb unloaded with or without brief weightbearing (HU±SC, n=14), and hindlimb unloaded with brief weightbearing and 90 Hz oscillations (HU+90, n=7) rats. Low-intensity vibrations prevented a decoupling of the relationship observed in hindlimb unloaded rats. *: significant regression; p <0.05

Age-Related Changes

Baseline IVD height of AC rats was 0.79±0.05 mm and did not change over the 4 wk, but nucleus pulposus area of AC rats was 25±10% greater (p=0.023) than that of BC (Figures 1 and 2). For mechanical structural properties, torsional stiffness K of AC discs was greater than of BC by 46±17% (p=0.016) (Table 1). During creep, the short time constant τ₁ of AC discs was longer by 3.6±1.3s (p=0.017) (Table 2). No age-related differences were measured in creep properties (p≥0.262), torsional properties (p≥0.085), axial stiffness (p≥0.143), and modulus properties (p≥0.115) of the L4-L5 motion segment (Tables 1 and 2, Figure 3). Total lumbar muscle volume of the AC rats increased over 4 wk of aging (p=0.003) but individual psoas or paraspinal muscle area did not change significantly (p≥0.086) (Figure 4, Tables 3–4).

Table 4.

Baseline psoas and paraspinal muscle area of AC, HU±SC and HU+90 rats [cm²]

Muscle-Location	AC	HU±SC	HU+90
Psoas-L4	0.62±0.04	0.65±0.05	0.68±0.05
Psoas-L5	0.84±0.09	0.81±0.07	0.84±0.07
Paraspinal-L4	2.67±0.18	2.67±0.16	2.63±0.19
Paraspinal-L5	2.57±0.10	2.39±0.24	2.44±0.24

Values are presented as mean±SD; n=7 for age-matched (AC) and hindlimb unloaded with brief 90 Hz oscillations (HU+90); n=14 for hindlimb unloaded with or without brief weightbearing (HU±SC)

DISCUSSION

The aim of this study was to assess the degradation of lumbar IVD biomechanics/morphology and the surrounding spinal muscle morphology of hindlimb unloaded rats and to evaluate the ability of low intensity vibrations to mitigate these changes. Four weeks of hindlimb unloading hypotrophied the disc, deconditioned the spinal segment and altered the axial and torsional mechanical behavior of the disc with almost no impact from brief interruptions of bipedal posture in sham controls (SC). Superimposing 90 Hz, low-magnitude mechanical oscillations on brief periods of upright weight bearing maintained the morphology of the IVD and psoas muscle but not paraspinal muscle, and partially countered the disruption of the axial, but not torsional, mechanics of the L4-L5 segment during hindlimb unloading. Thus, the incorporation of brief periods of low-intensity, high-frequency mechanical signals to weight bearing activities may not only provide morphological and biochemical benefits to the IVD as previously shown [21], but can also positively impact biomechanical function during long periods of spinal deconditioning.

Previously, we demonstrated that unloading is detrimental to the glycosaminoglycan content of rat IVD [21]. Using a similar experimental design, here, we assessed changes to the geometry and axial mechanics of the disc. Unloaded discs had a smaller nucleus pulposus area and displayed a weaker axial NZ modulus, a measure of segmental motion from the neutral position with particular sensitivity to degeneration [31] induced by altered glycosaminoglycan content [26]. Discs of hindlimb unloaded rats also had a greater compressive range of motion, modulus and instantaneous stiffness than of age-matched animals. These unloading-induced changes to the compressive properties of the disc were similar to the greater spinal compressibility of bed rest subjects (49% vs 59%) [3] and stiffening of rodent discs following five days of microgravity (25% vs 26%)[17].

The ability of 90 Hz vibrations to fully or partially maintain compression-tension and creep properties of the disc during unloading may be associated with differences in biochemical and morphologic properties between vibrated and control rats. Under normal loading conditions, the incompressible fluid of the nucleus facilitates the transmission of axial forces into circumferential tension and axial deformation [38]. However, discs with less nuclear glycosaminoglycan do not properly engage the inner annular fibers during axial compression and overstress the outer annular fibers in the radial and axial direction [32], leading to increased stiffness. In support of this association between mechanical, structural, and biochemical properties in the IVD subjected to unloading and vibrations, we recently showed that 90 Hz oscillatory motions superimposed on intermittent periods of squat-like posture retain the area and glycosaminoglycan content in the nucleus pulposus of hindlimb unloaded rats [21]. Thus, application of the 90 Hz mechanical signal during altered loading conditions may preserve the effective load transfer through the disc.

Daily application of 90 Hz vibrations during brief periods of erect weight bearing also prevented the uncoupling of disc mechanics and morphology during hindlimb unloading. The negative relationship between NZ modulus and change in posterior disc height of normally ambulating rats may represent the mechanical relaxation with increasing posterior height. The posterior region of the IVD is of note for its sensitivity to damage [7] and degeneration [13], and its greater morphological change during bed rest [11]. Hindlimb unloading eliminated the mechanical-morphological relationship, indicating a disruption of disc constituents and their ability to maintain normal function. While the application of 90 Hz vibrations maintained the mechanical-morphological relationship of HU+90 discs, it did not maintain the posterior height of hindlimb unloaded IVD [21].

Spinal muscular loading is a principal source of axial loading of the rodent spine. Therefore, changes (and differences) in spinal muscle area may provide insight into the mechanical loading environment of discs following the absence of dynamic loading. After four weeks, spinal muscle volume increased similarly in all groups. When stratified into specific muscle groups, hindlimb unloaded rats had increased psoas muscle area and decreased paraspinal muscle area at L4, similar to bed rest studies [2,11,34]. The lost paraspinal area and gained psoas area perhaps reflects the need to recruit alternate muscle fiber types. For example, bed rest may reduce the area of postural muscles, like the paraspinal muscles, by replacing large area type II fibers with small type I fibers [35]. By contrast, the increase in psoas muscle area with hindlimb unloading may derive from a shift of fiber type as occurs with the disc degenerative procedure of limb amputation [36].

Although the consequences of reduced dynamic loading on the musculature remain unclear, physical countermeasures may target paraspinal muscles and translate into mechanical benefits to the IVD. Here, the application of high-frequency (90 Hz), low-intensity (0.2 g) vibrations in an upright posture did not protect the paraspinal muscle but prevented changes to the psoas muscle area. In a previous bed rest study [20,34], the application of high-frequency (30 Hz), low-intensity (0.3, 0.5 g) vibrations superimposed on an axial force (0.6 body weight) did not prevent paraspinal atrophy or affect psoas muscle area. By contrast, resistive exercise with and without high-frequency (16–26 Hz), low-intensity (0.7 g) vibrations superimposed on a larger axial force (1.2–1.8 body weight) countered paraspinal muscle atrophy during prolonged bed rest but was largely ineffective at mitigating increased psoas muscle [19,37]. While both psoas and paraspinal muscles are involved with spinal compression, paraspinal muscles are additionally involved in torso rotation [38]. Because paraspinal muscle area was highly related (R²=0.99) to a torsional property of the disc, muscle contractions from the extremely brief bouts of postural stabilization (15 min/day), rather than weightbearing, may have maintained the torsional stiffness during unloading.

While the hindlimb unloaded rat served to simulate reduced spinal loading, the cellular and functional disparities need to be kept in mind when interpreting the consequences of any applied load. Despite the limitations associated with using a quadruped to simulate the effects of unloading and loading human IVD, when normalized by the geometry, biomechanical properties of rodent lumbar IVD are similar to those of human IVD [25,27,39]. Moreover, upright posture in humans induces discal bending with the anterior region of the IVD loaded in tension and the posterior region in compression [40]. Surprisingly, a rat disc may be loaded in a similar fashion as suggested by thinner collagen fibers suited for recoverable creep in the posterior region [4]. While hindlimb unloading produces high tension on the tail, the induced mechanical environment of the lumbar disc remains unclear and may account for the disparity noted between the hypotrophy of IVD with hindlimb unloading [22] and hypertrophy with human spinal unloading [11]. Furthermore, the relatively young age of the rodents may have influenced the response to loading and it is entirely possible that older rodents and humans respond differently. Lastly, while CT is commonly used to quantify muscle volume and disc height, MRI and histology will be required to further characterize the structural and functional changes induced by unloading and the vibratory intervention.

Low-level, high-frequency mechanical signals partially maintained the function and morphology of a spinal motion segment, and prevented the loss in the relationship between IVD mechanics and morphology during spinal deconditioning. Clinically, vibrations mitigate the increased incidence of back pain and normalize IVD morphology during prolonged bed rest and suggest a putative link [2]. The inability of the signal to fully protect the IVD and muscle from the detrimental effects of immobilization may suggest that efficacy could be improved by the incorporation of complementary mechanical stimuli [19,41]. However, the complimentary stimulus must be applied with caution to not exacerbate pain [19] since deconditioned discs are mechanically vulnerable to damage [7]. Further, it is possible that contractions from the spinal muscles indirectly influenced maintenance of the IVD [42]. Nevertheless, the results highlight the fundamental physiologic relevance of low-level mechanical signals at high frequencies to maintain mechanical tissue function during deconditioning [43]. Whether vibratory signals can be used clinically to prevent or treat muscle atrophy, back pain or disc degeneration during conditions of reduced dynamic loading or immobilization requires further investigation.