Multi-laminate Annulus Fibrosus Repair Scaffold with an Interlamellar Matrix Enhances Impact Resistance, Prevents Herniation and Assists in Restoring Spinal Kinematics

Abstract

Focal defects in the annulus fibrosus (AF) of the intervertebral disc (IVD) arising from herniation have detrimental impacts on the IVD’s mechanical function. Thus, biomimetic-based repair strategies must restore the mechanical integrity of the AF to help support and restore native spinal loading and motion. Accordingly, an annulus fibrosus repair patch (AFRP); a collagen-based multi-laminate scaffold with an angle-ply architecture has been previously developed, which demonstrates similar mechanical properties to native outer AF (oAF). To further enhance the mimetic nature of the AFRP, interlamellar (ILM) glycosaminoglycan (GAG) was incorporated into the scaffolds. The ability of the scaffolds to withstand simulated impact loading and resist herniation of native IVD tissue while contributing to the restoration of spinal kinematics were assessed separately. The results demonstrate that incorporation of a GAG-based ILM significantly increased (p<0.001) the impact strength of the AFRP (2.57 ± 0.04 MPa) compared to scaffolds without (1.51 ± 0.13 MPa). Additionally, repair of injured functional spinal units (FSUs) with an AFRP in combination with sequestering native NP tissue and a full-thickness AF tissue plug enabled the restoration of creep displacement (p=0.134), short-term viscous damping coefficient (p=0.538), the long-term viscous (p=0.058) and elastic (p=0.751) damping coefficients, axial neutral zone (p=0.908), and axial range of motion (p=0.476) to an intact state. Lastly, the AFRP scaffolds were able to prevent native IVD tissue herniation upon application of supraphysiologic loads (5.28 ± 1.24 MPa). Together, these results suggest that the AFRP has the strength to sequester native NP and AF tissue and/or implants, and thus, can be used in a composite repair strategy for IVDs with focal annular defects thereby assisting in the restoration of spinal kinematics.

Graphical Abstract

1. Introduction

Intervertebral discs (IVDs) consist of a gelatinous core known as the nucleus pulposus (NP), which is circumferentially constrained by a fiber-reinforced, multi-laminate structure known as the annulus fibrosus (AF). The NP’s hydrophilic extracellular matrix (ECM) is composed predominantly of aggrecan and collagen type II. The AF consists of 15–25 concentric lamellae composed of type I collagen fibers aligned in a ± 28–43° angle-ply microarchitecture in the outer AF (oAF), which gradually transitions to a less organized inner AF (iAF) containing glycosaminoglycan (GAG) and type II collagen. (Urban and Roberts, 2003) Between each oAF lamellae, non-fibrillar ground substance (e.g., glycosaminoglycan: GAG), elastin, collagens type I and VI, and cells form an interlamellar matrix (ILM). (Tavakoli et al., 2016) The highly organized, fiber-reinforced structure of the oAF resists tension and torsion during spinal bending and twisting but also bears hoop stresses developed from pressurization of the NP during compressive loading of the IVD. (Ducheyne, P. Healy, 2011; Long et al., 2016; Patwardhan et al., 2003; Schultz et al., 1982) Furthermore, the ILM reinforces the oAF against radially-directed intradiscal pressures (IDPs) and prevents de-lamination of adjacent lamellae. (Adam et al., 2015) Thus, the ECM composition and organization of these IVD tissues allow for stable, multi-axial motion of the spine through their mechanical interplay.

IVD herniation (IVDH) occurs when the NP protrudes or extrudes from its annular confines, which is often the result of abrupt loading (i.e. impact loading) consisting of bending and twisting. As a result, nerve roots become compressed and irritated causing significant pain and disability. Discectomy is often required to remove the extruding NP tissue with approximately 500,000 discectomies performed annually in the United States (Figure 1). (Miller et al., 2018; Mirza et al., 2006) Although the oAF can heal via fibrotic scarring, (Torre et al., 2018) the resulting unorganized tissue does not restore native oAF micro-architecture or strength leading to re-herniation in 9–25% of patients at 4- and 10-years post-discectomy. (Atlas et al., 2005; Hu et al., 1997) Additionally, it has been noted that patients who have less aggressive discectomies (i.e. less NP tissue removed) demonstrate improved maintenance of IVD height, biomechanics, and reduced pain. However, these patients are at increased risk for recurrent herniation, (Bailey et al., 2013; Lequin et al., 2012) presumably due to having increased pressurization upon re-loading the IVD because less NP tissue is removed during the surgery.

Figure 1: Schematic depicting the treatment of disc herniation via either minimal or aggressive discectomy procedures to restore mechanical integrity and assist in the support and restoration of native spinal loading and motion.

Both approaches would require an AF replacement biomaterial and an AF outer defect closure; however, an aggressive approach could also require the excised NP tissue to be replaced with an NP surrogate.

Taken together, a need exists to develop mechanically competent AF repair biomaterials that can immediately close focal AF defects caused by herniated NP that mimic the native structure and function of the oAF. (Sharifi et al., 2014) Moreover, these biomaterials should resist impact loading without herniation, and thus, contribute to the restoration of functional spinal unit kinematics (FSU). The later may be achieved by either sequestering1) residual native NP tissue following a less aggressive discectomy, or 2) an NP replacement implanted either during the index discectomy procedure or at a later time-point following the further expected degeneration of the IVD (Figure 1). (Elliott et al., 2006; Fountas et al., 2004; O’Connell et al., 2011) In either case, the AF repair must be strong enough to restrain the intradiscal pressurization of the native NP or NP surrogate and secure any residual AF tissue to afford the best opportunity to restore FSU kinematics while potentially mitigating progressive IVD degeneration.

In response to this need, several groups have developed biomaterial scaffolds for repairing and regenerating focal AF defects. Innovative injectable adhesives have been formulated to completely fill the defect pathway that remains in the iAF and oAF following discectomy. These include hydrogels composed of genipin-crosslinked fibrin (Cruz et al., 2018; Frauchiger et al., 2018; Long et al., 2016) and ultraviolet light curable riboflavin cross-linked collagen. (Pennicooke et al., 2018; Sloan et al., 2017) Current challenges with these biomaterials include the need to improve adhesive strength to resist reherniation upon application of physiologic spinal loads and NP pressurization. (DiStefano et al., 2019; Sloan et al., 2019) Others have investigated composite repair strategies for treating focal AF defects using pre-formed, full thickness AF scaffold plugs made of polymer secured within the defect pathway using auxiliary synthetic membranes/patches affixed via suture to the oAF. (Long et al., 2017; Pirvu et al., 2015) Results from physiologic kinematic testing of these approaches illustrated that implant herniation occurred in approximately 80% of the specimens due in part to patch failure, and an inability of the repair to mimic the anisotropic tensile properties of the native oAF. (Long et al., 2017) Conversely, others have demonstrated successful retention of AF scaffold plugs within IVD models using an oAF patch; however, compressive stresses experienced by the IVD were at the low end of the physiologic spectrum. (Pirvu et al., 2015) Thus, these assessments may have not adequately evaluated the mechanical strength of the AF repairs required for successful use in vivo.

To address these limitations, our group has previously reported on the development and characterization of a collagen-based, multi-laminate, angle-ply AF repair patch (AFRP). (Borem et al., 2017; McGuire et al., 2016) It is envisioned that these biomimetic oAF patches can be used in conjunction with biomaterials that replace the NP and fill the AF defect pathway to prevent herniation following discectomy repair. We have demonstrated the biomimetic nature of the AFRP with respect to oAF micro-architecture, static and dynamic tensile material properties, and their ability to sequester an NP replacement which together partially restored FSU kinematics following repair of injured IVDs. (Borem et al., 2017) We have also demonstrated that attachment of the AFRP to adjacent oAF tissue can be achieved via 4–0 FiberWire suture, which prevented patch detachment following application of supraphysiologic loads to repaired FSUs. (Borem et al., 2017) However, prior investigations also demonstrated that radially-directed impact burst strength of the AFRP was within the mid-range of physiologic pressures experienced by the IVD. Moreover, the AFRP’s ability to sequester fully pressurized native NP and iAF tissue within an injured IVD experiencing physiologic loading has yet to be assessed. Herein, we hypothesized that inclusion of a glycosaminoglycan (GAG)-based ILM between the layers of the AFRP would significantly increase its impact burst strength. Moreover, it was thought that these implants would demonstrate the strength required to resist herniation of fully pressurized native NP and AF tissue within the injured IVD; thus, contributing to the restoration of FSU kinematics. Therefore, the broad objective of the work herein was to further demonstrate the mechanical competency of the AFRP.

2. Materials and Methods

2.1. FABRICATION OF ANNULUS FIBROSUS REPAIR PATCHES (AFRPs)

Multi-laminate “angle-ply” AFRPs were developed and assembled from decellularized porcine pericardium to form tri-layer “3-layered” scaffolds and crosslinked in 6mM 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC) with 1.2mM N-hydroxysuccinimide (NHS) as previously described. (Borem et al., 2017; McGuire et al., 2016) Briefly, the pericardium’s fiber orientation of individual decellularized pericardium sheets was laid on top of a light box and sections of the tissues that clearly defined collagen fiber-preferred direction/alignment were identified in the fibrous pericardium and cut into squares. The fiber alignment of each individual square was then oriented ±30° (verified via a protractor) relative to a stationary grid containing a common horizontal axis. Once aligned the multi-laminate sheets were placed upon a dissolvable embroidery backing material (Sulky Fabri-Solvy; 100% polyvinyl alcohol), which allowed for easy positioning within a sewing machine and enabled sewing needle penetration throughout all pericardium layers. A square pattern was sewn around the periphery of the of the sheets using a thread diameter equivalent to a 2–0 suture followed by removal of excess tissue and backing material. AFRPs were maintained in a storage solution containing phosphate buffered saline (PBS) and protease inhibitor (Sigma: S8820) at 4°C prior to testing. A total of n=20 AFRPs were evaluated separately for this study: histological analysis (n=5), impact resistance testing (n=15), functional spinal unit biomechanical testing (n=5).

2.2. HA-GEL PREPARATION AND AFRP ENRICHMENT

Hyaluronic acid (HA)-gel was prepared by reconstituting 1% sodium hyaluronate (a water-soluble salt form of HA) in PBS. AFRPs were then enriched by injecting the HA-gel between the AFRP lamellae (~62.5ug/mg) to form a GAG-based inter-lamellar layer (Figure 2A&B). The injection procedure consisted of a 28-gauge needle inserted at the edge of the suture line of the AFRP. The transparency of an AFRP layer allows for the visualization of the needlepoint. The HA was then injected between the adjacent layers and was visually confirmed via minimal inflation of the AFRP.

Figure 2: Glycosaminoglycan (GAG) enrichment of the AFRP.

A) Representative schematic illustrating the location of the proteoglycan-rich interlamellar matrix (purple) found between native AF lamellae. B) Injection of hyaluronic acid (HA) between the AFRP lamellae. C) Representative image depicting Alcian blue staining confirming the incorporation GAG between the AFRP lamellae [* indicates AFRP layer; ^ indicates positively stained GAG] (insert: a magnified image of GAG-enriched AFRP).

2.3. HISTOLOGICAL STAINING FOR CONFIRMATION OF INTER-LAMELLAR GAG

AFRPs were fixed in 10% neutral-buffered formalin for 24 hours before undergoing successive washes in graded ethanol, xylene, and paraffin followed by embedding and sectioning to 5μm thickness. Slides were stained with Alcian Blue (1% Alcian Blue in 3% aqueous acetic acid; pH 2.5) and counterstained with 0.1% aqueous Nuclear Fast Red for visualization of glycosaminoglycan deposition to confirm incorporation between the AFRP’s lamellae. Histological images were then captured on a Zeiss Axio Vert.A1 microscope with AxioVision software (SE64 Rel. 4.9.1 SP08–2013).

2.4. IMPACT RESISTANCE TESTING

Radial impact strength testing was performed on individual AFRP samples according to methods previously described, and was modeled after ASTM D1709: “Standard Test Methods for Impact Resistance of Plastic Film by the Free-Falling Dart Method” with minor modification. (Borem et al., 2017) Briefly, representative samples of crosslinked and crosslinked + HA-gel enriched AFRPs (3-ply; n=15/group; AFRP dimensions: 12 mm (L) × 12 mm (W) × 0.75mm (T)) were tested using a custom designed free-fall impact testing drop-tower. The base platform of the drop-tower consisted of a tissue-holding clamp and four vertical rails, which guided a free-falling platform. The tissue holding apparatus consisted of two stacked blocks lined with coarse-grit sandpaper each having an aligned thru-hole of 6.25mm diameter. An AFRP was interposed between the two blocks centered over the two thru-holes. Subsequently, a 6mm diameter steel ball attached to a 3-inch pushrod was placed in contact with the AFRP via the thru-hole in the superior block. Various weights ranging from 0.32–0.65 kg were stacked on the free-fall platform, which was then dropped from a constant height of 0.178 m. Impact energy (E) was calculated using the equation for kinetic energy, $E=\frac{1}{2}*m*v^2$, where m = mass (variable) and v = velocity (1.867 m/s). Velocity was calculated as $\sqrt{2*g*h}$, where g = gravity (9.807 m/s²) and h = height (0.178 m). The resultant ball-burst pressure was calculated given the maximum force at rupture and its relationship with ball-burst pressure and geometric constraints according to established equations (Equations 1–3) given the geometric constraints of our test set-up (Freytes et al., 2005; McGuire et al., 2016).

$$P=\frac{F}{A}=\frac{F}{(2\pi d^2(1-cos(\phi ))}$$ Equation 1:

$$\phi =\pi -\left(\frac{\pi }{2}-ta{n}^{-1}\left(\frac{d}{f}\right)\right)-ta{n}^{-1}\frac{a}{b}$$ Equation 2:

$$f=\sqrt{(b^2}+a^2-d^2)$$ Equation 3:

Where P is the ball burst pressure, F is the maximum recorded burst force, A is the contact area between the patch and the surface of the steel ball, φ is the contact angle between the AF patch material and ball, d is the radius of the steel ball, f is the magnitude of the vector representing the stretching material which is geometrically determined from: a; the distance between the central axis of the ball and tissue clamp set-up (3.25mm), b; the position of the steel ball and push rod relative to its starting point (3mm) which maintains the tangential relationship between the patch material and surface of the ball as described by (Crapo et al., 2011; Freytes et al., 2005; McGuire et al., 2016)

2.5. PREPARATION OF FUNCTIONAL SPINAL UNITS

Bovine caudal tails, from 2–3-year-old calves, were obtained from a local abattoir and transported on wet ice to the lab within an hour. Excess tissue and ligaments surrounding the vertebral bodies and intervertebral discs were removed via dissection, and functional spinal units (FSUs: vertebrae-IVD-vertebrae) with posterior elements attached were isolated via a bandsaw. FSUs were harvested from two caudal levels (cc1–2 to cc2–3) and were potted using 3mm steel rods and urethane potting resin to prevent slippage of the samples during testing. Prior to testing, FSUs were wrapped in gauze saturated with PBS + protease inhibitor and stored at −80°C. Samples were thawed within the sealed zip-lock bag, which was submerged for four hours in PBS + protease inhibitor at ambient temperature.

2.6. BIOMECHANICAL EVALUATION OF FUNCTIONAL SPINAL UNIT AXIAL AND TORSIONAL KINEMATICS

FSU’s (n=5/group) underwent repeated measures kinematic testing protocol according to methods previously described by our group, with minor modification. (Borem et al., 2017) Briefly, FSU’s (n=5/group) were tested with a repeated measure protocol using the following groups: Intact, Annulotomy, and Repair (Figure 3A). To initiate injury of the FSU, an annulotomy was performed by perforation of the IVD using a 6mm diameter biopsy punch (7mm depth) and subsequent removal of the AF tissue. A 6mm diameter defect was chosen to represent a “worst-case” scenario which would be required to implant an NPR, as observed in accordance with previous literature. (Heuer et al., 2008; Wilke et al., 2006) The annulotomy group consisted of the 6mm diameter annular defect, and the excised tissue was marked with a tissue marker for orientation and stored in PBS + protease inhibitor. Additionally, the outer annular defect was closed with an HA-gel enriched AFRP to prevent herniation of native NP material, and to ensure that subsequent testing with the AF tissue plug was performed with a fully pressurized NP. Of note, attachment of the AFRP to the outer AF did prevent overt herniation of the native NP, as bulging of NP tissue into the AFRP was noted indicating radial migration. The repair group consisted of filling the full-thickness annular defect using the previously excised intact native AF tissue (AF tissue plug) recovered during the annulotomy, and this tissue was secured in place with an HA-gel enriched AFRP. The excised AF tissue was inserted with the inner AF portion of the AF lamellae implanted first into the defect, and the plug was then gently pushed into place with minimal force to become flush with the native outer AF layers prior to being secured by the AFRP. AFRP’s (dimensions: 7mm (L) × 7mm (W) × 0.75mm (T)), were secured to the IVD via a 4–0 FiberWire suture. Suturing of the AFRP consisted of suturing at the four corners with sutures being passed through the AFRP 1–2mm from the edge of the patch, and suture throws were made in alignment with the ±30° collagen fibrils of the AFRP.

Figure 3: Study design for in situ kinematic testing of bovine IVD functional spinal units (FSUs).

A) Representative images of the progression of sample testing (intact, annulotomy, and repaired groups) for the kinematic evaluation of FSUs. B) Loading scheme for FSU testing depicting creep loading, axial cyclic tension-compression loading, torsional loading, and slow constant-rate ramp loading. C-E.1) Representative graphs depicting the axial, creep, and torsional response of an IVD and its associated parameter measures. E.2) Graph depicting the torsional hysteresis height and width; parameters were calculated based on the dimensions of their respective geometric axes (orange solid line = height; green dotted line = width).

The kinematic testing scheme and loading conditions used herein were performed in accordance with previous literature with minor modification, which has been used to evaluate AF’s repair biomaterials. (Borem et al., 2017; Cruz et al., 2018; Johannessen et al., 2006; Likhitpanichkul et al., 2014; Long et al., 2016) FSU’s first underwent creep loading on a Bose ElectroForce (model: 3220, TA Instruments) equipped with a 100-lb. load cell and a test chamber filled with 1xPBS/protease inhibitor at 25°C. Samples were loaded to a mean amplitude level of −0.125 MPa in compression and then underwent a 1-hr. creep period at −0.50 MPa in compression and then returned to a mean amplitude level of −0.125 MPa in compression prior to removal. Samples were then directly transferred to a servohydraulic test frame (model: 8874, Instron) fitted with a 20kN load cell, and a mean amplitude load of −0.125 MPa in compression was applied immediately. Post-analysis was performed to confirm the FSUs disc height remained consistent between testing frames. Samples were then subjected to 35 cycles of axial compression (−0.50 MPa) and tension (0.25 MPa) at 0.1 Hz. Compression was then maintained at −0.50 MPa in compression as samples underwent 35 torsional cycles to ±3°. Finally, samples underwent a slow-rate compressive ramp (1 N/s) from a mean amplitude level of −0.125 MPa to −0.50 MPa in compression (Figure 3B). Between testing groups, FSUs were wrapped in a PBS + protease inhibitor saturated gauze and placed within a polyethylene-linear low-density bag (Ziplock bag) and rested under no mechanical load for an eighteen-hour rest period to permit tissue equilibration. Tensile and compressive stiffness was determined using a linear fit of the loading force-displacement curve from 60–100% of the 35^th cycle (Figure 3C). (Borem et al., 2017; Likhitpanichkul et al., 2014) The axial range of motion was defined as the total peak-to-peak displacement of the IVD, and the axial neutral zone length was determined by fitting a third-order polynomial to the data and finding the maxima and minima with the correlating range between the peaks. (Borem et al., 2017; Johannessen et al., 2006) A non-linear constitutive model was fit to the creep data (Figure 3D) using GraphPad Prism 7 software to yield elastic (Ψ) and viscous (η) damping coefficients for the short-term (η₁ and Ψ₁) and long-term (η₂ and Ψ₂), as described previously. (Borem et al., 2017; Johannessen et al., 2006) Torsional stiffness was calculated from a linear fit of the loading torque-rotation curve of the 35th cycle. Torque range and axial range of motion (RoM) was calculated as the peak-to-peak torque and displacement, respectively. (Borem et al., 2017; Likhitpanichkul et al., 2014) Changes in torque hysteresis were calculated by measuring the height and width of the curve along the geometric axes as shown in Figure 3E&F. (Bezci et al., 2018) The constant-rate slow-ramp compression stiffness was determined using a linear fit of the slow-ramp load-displacement response.

2.7. COMPRESSION TO FAILURE OF REPAIRED FSUS

Following kinematic testing of FSUs, repaired samples were loaded to a mean amplitude of −0.125 MPa in compression and then compressed at a rate of 5 mm/min until FSU failure (i.e., 1) NP tissue herniated, 2) AFRP failure, or 3) the load was no longer being applied axially (due to the convex end-plates coming into contact with one another), and the FSU began to initiate bending). Ultimate compressive strength was calculated from the sample load and the initial cross-sectional area of the IVDs. Digital video was used to record mode and time of failure.

2.8. STATISTICS

Statistical analysis was performed on raw data using GraphPad Prism 7 software. Results are represented as mean ± standard deviation (SD), and significance was defined as (p≤0.05). Impact resistance data was evaluated using a one-way ANOVA with Dunnett’s post-hoc analysis. Kinematic data were evaluated using a one-way repeated measures ANOVA followed by Dunnett’s post-hoc analysis.

3. Results

3.1. CONFIRMATION OF GAG-ENRICHMENT OF AFRPs

HA-gel enriched AFRPs were histologically evaluated to confirm the presence of GAG between the AFRP lamellae following injection. Positive Alcian blue staining of the crosslinked + HA-gel AFRPs illustrated positive staining within the interlamellar region confirming the incorporation of GAG between the AFRP lamellae (Figure 2C).

3.2. IMPACT TESTING

Multi-laminate AFRPs underwent radially-directed impact loading to evaluate their resistance to failure during impact loading (Figure 4A&B). Mean impact strength of noncrosslinked, crosslinked, and crosslinked + HA-gel AFRPs was 1.51 ± 0.13 MPa, 2.04 ± 0.06 MPa, 2.57 ± 0.04 MPa, respectively (Figure 4C). Additionally, the average kinetic energy absorbed by non-crosslinked, crosslinked, and crosslinked + HA-gel AFRPs was 0.51 ± 0.04 J, 0.69 ± 0.02 J, 0.87 ± 0.01 J, respectively (Figure 4D). Expectedly, impact strength was significantly increased following crosslinking (p<0.001) and crosslinking + HA-gel (p<0.001) compared to non-crosslinked AFRPs. Furthermore, impact energy was significantly increased in crosslinked AFRPs (p<0.001) and crosslinked + HA-gel AFRPs (p<0.001) compared to non-crosslinked AFRPs.

Figure 4: Biomechanical evaluations of non-crosslinked, crosslinked, and crosslinked + HA-gel AFRPs subjected to radially-directed impact pressures.

A) Representative images of a custom designed free-fall impact testing drop-tower consisting of a base platform with a vertical rail system to guide a free-falling platform with a predetermined weight. B) Representative image of the AFRP prior to testing and impact failure mechanism. C) Representative schematic illustrating the radial loading axis the NP exerts on the AF through increases in intradiscal pressures (IDPs). D) Graph depicting the average radial impact strength withstood by AFRP testing groups. The dotted lines indicate reported human values of IDPs 0.1–2.3 MPa. E) Graph depicting the average radial impact kinetic energy absorbed by the AFRP testing groups. Solid lines connecting groups indicates a significant difference (p<0.05). Results are depicted as mean ± SD.

3.3. KINEMATIC RESULTS

Axial and torsional FSU kinematics were evaluated to determine the effect of an isolated AF defect (annulotomy) and subsequent repair using the full-thickness AF tissue plug (excised during annulotomy) secured in place by an AFRP. Pressurization of the IVD due to the intact NP was confirmed as the AF tissue plug immediately extruded/herniated prior to the application of significant mechanical loading without sequestration by the AFRP (Figure 5).

Figure 5: Pressurization of the IVD prevented the repair of a native AF Plug only group without an outer AFRP closure.

A) Representative image of an annular defect following annulotomy procedure (purple dotted circle). B) Representative top view of the excised AF tissue plug being implanted into FSU with minimal force (red dotted circle). C) Representative side view illustrating the immediate extrusion of the AF tissue plug out of the annular defect immediately following implantation and application of loading (blue dotted circle).

3.3.1. CREEP LOADING

Creep loading of annulotomized and repaired FSUs did not significantly alter the step displacement or short-term elastic damping coefficient (Figure 6A&B). However, annulotomized FSUs resulted in significant increases in creep displacement (Annulotomy: 1.29 ± 0.17 mm, Intact: 1.56 ± 0.16 mm; p=0.03), long-term elastic damping coefficient (Annulotomy: Ψ₂: 201.30 ± 33.48 N/mm, Intact: 140.10 ± 7.90 N/mm; p=0.022), and short-term and long-term viscous damping coefficients (Annulotomy: η₁: 7532.40 ± 2.807.40 × 10³ Ns/mm and η₂: 46.59 × 10⁴ ± 7.36 × 10⁴ Ns/mm, respectively, Intact: η₁: 4.73 × 10³ ± 2.38 × 10³ Ns/mm and η₂: 3.70 × 10⁵ ± 2.42 × 10⁴ Ns/mm, respectively; p₁=0.003 and p₂=0.024) compared to intact controls (Figure 6C–F). However, repair with the native AF tissue plug sequestered by an AFRP, restored the creep displacement (1.74 ± 0.13 mm; p=0.134), long-term elastic damping coefficient (Ψ₂: 144.48 ± 16.77 N/mm; p=0.751), and short- and long-term viscous damping coefficient (η₁: 5839.60 ± 3.01 × 10³ Ns/mm and η₂: 31.33 × 10⁴ ± 4.28 × 10⁴ Ns/mm, respectively; p₁=0.538 and p₂=0.058) to intact levels.

Figure 6: Creep kinematic testing results of bovine IVD FSUs.

Graphs depicting the A) step displacement, B-C) short-term elastic and viscous damping coefficients, D) creep displacement, E-F) long-term elastic and viscous damping coefficients. Solid lines connecting groups indicates a significant difference (p<0.05). Results are depicted as mean ± SD.

3.3.2. AXIAL CYCLIC LOADING

Axial cyclic kinematic testing of annulotomized and repaired FSUs did not significantly alter cyclic axial compressive and tensile stiffness, nor slow ramp compressive stiffness parameters compared to intact controls (Figure 7A&B). Conversely, annulotomized FSUs resulted in a significant increase in axial range of motion (Annulotomy: 4.03 ± 0.60 mm, Intact: 3.45 ± 0.64 mm; p=0.045) and neutral zone (Annulotomy: 1.55 ± 0.25 mm, Intact: 1.28 ± 0.29 mm; p=0.037) parameters compared to intact controls (Figure 7C&D). However, repair of the injured IVD with the native AF tissue plug sequestered by the AFRP (repair), restored these parameters to intact values (axial range of motion: 3.67 ± 0.64 mm; p=0.476 and neutral zone: 1.31 ± 0.25 mm; p=0.908).

Figure 7: Axial cyclic tension-compression kinematic testing results of bovine IVD FSUs.

Graphs depicting the A) axial compressive stiffness, B) slow ramp compressive stiffness, C) tensile stiffness, D) axial range of motion, E) axial neutral zone. Solid lines connecting groups indicates a significant difference (p<0.05). Results are depicted as mean ± SD.

3.3.3. TORSIONAL CYCLIC LOADING

Focal injury to the AF by annulotomy resulted in significant changes in FSU torsional kinematics, which were not restored with the repair methodology (Figure 8). Torsional rotation of annulotomy and repaired FSUs demonstrated a significant decrease in torsional stiffness (0.48 ± 0.05 Nm/°: p=0.001 and 0.53 ± 0.06 Nm/°: p=0.024, respectively) (Figure 8A), torque range (3.45 ± 0.27 Nm: p=0.0002 and 3.68 ± 0.37 Nm: p=0.022, respectively) (Figure 8B), and torque hysteresis height (0.28 ± 0.08 Nm: p=0.028 and 0.30 ± 0.09 Nm: p=0.023, respectively) (Figure 8C) compared to intact FSUs (0.66 ± 0.05 Nm/°, 4.60 ± 0.23 Nm, 0.42 ± 0.10 Nm, respectively).

Figure 8: Torsional kinematic testing results of bovine IVD FSUs.

Graphs depicting A) torsional stiffness, B) torque range, C) torque hysteresis height, D) torque hysteresis width. Solid lines connecting groups indicates a significant difference (p<0.05). Results are depicted as mean ± SD.

3.3.4. COMPRESSION TO FAILURE

Axial compressive testing of annulotomy and repaired FSUs were assessed for their failure strength concomitant with visual assessment for herniation or extrusion (Figure 9). Compression to failure testing demonstrated no failure in the AFRP attachment nor herniation of native IVD tissues. Furthermore, no statistical differences in ultimate compressive strength were observed between annulotomy and repair groups (Annulotomy: 5.28 ± 1.24 MPa; Repair: 3.84 ± 1.05 MPa; p=0.101) or compressive displacement (Annulotomy: 2.58 ± 0.50 mm; Repair: 2.96 ± 1.37 mm; p=0.619).

Figure 9: Compression to failure testing results of bovine IVD FSUs.

Graphs depicting compression to failure A) strength and B) displacement. Representative images obtained during failure testing depicting: C) The FSU was loaded to an initial mean amplitude level of −0.125 MPa in compression prior to compressive failure testing. D) Representative image depicting compressive failure of the annulotomy group, which did not demonstrate herniation of the native IVD tissues. E) Representative image depicting compressive failure of the repaired group, which did not demonstrate herniation, but did result in radial bulging of the AFRP. Results are depicted as mean ± SD.

4. Discussion

The primary goal of successful AF repair following discectomy is to mitigate progressive IVD degeneration by preventing recurrent herniation of native IVD tissue or their replacements, and thus, help restore IVD biomechanics to an uninjured (intact) state. In the current study, the mechanical competency of an angle-ply oAF repair patch containing a GAG-based ILM was assessed for its ability to resist herniation following the application of simulated impact spinal loading and kinematic testing using an injured IVD model containing native NP and AF tissue. The primary findings from this study include: 1) incorporation of an ILM significantly increased the impact strength of the AFRP, and 2) the AFRP prevented herniation of pressurized native NP and AF tissue under applied physiological loading, thereby enabling the restoration of FSU axial kinematic parameters. Together, these results suggest the AFRP can be used as a mechanical closure system to sequester NP and AF tissues and/or repair implants within the IVD to assist in restoring its biomechanical function following discectomy.

The first significant finding from this study was that the inclusion of an ILM composed of an HA-hydrogel between the layers of the AFRP scaffold resulted in a significant increase in the patch’s impact strength. The ILM of the AF is composed of many different biomolecules that play a role in IVD mechanics. (Isaacs et al., 2014) Glycosaminoglycan, including chondroitin sulfate and HA, make up approximately 10–20% of the overall dry weight of the AF. (Perie et al., 2006; Roughley, 2004) Moreover, these molecules have been shown to play a functional role in contributing to the overall compressive properties of the IVD, aid the AF to resist IDP, and prevent delamination and shearing of lamellae. (Adam et al., 2015; Isaacs et al., 2014; Kirking et al., 2014; Nerurkar et al., 2011; Perie et al., 2006) Considering the important role of the ILM, researchers have developed tissue engineered AF scaffolds that contain ILM. (Nerurkar et al., 2009) This was achieved by culturing mesenchymal stem cells on multi-laminate electrospun scaffolds for 10–12 weeks, which eventually produced an ILM containing collagen and GAG that increased the mechanical properties of the constructs to near-native levels. (Nerurkar et al., 2009) Herein, we adopted a simplified (non-cellular-based) approach of injecting a GAG-based ILM surrogate within the AFRP and demonstrated that it significantly increased the impact strength of the construct compared to those without it. This is intuitive given the shock-absorbing role that GAGs play in other tissues. (Basalo et al., 2004; Lovekamp et al., 2006; Wilson et al., 2009) From a clinical perspective, the incorporation of the GAG-based ILM increased the AFRPs impact strength from approximately 3 g’s (non-crosslinked AFRPs (Borem et al., 2017)) to a conservative 10 g’s of gravitational force, which exceeds levels experienced by individuals when falling on their buttocks or riding rollercoasters. (Allen et al., 1994; Bibel, 2007) This suggests that the ILM-containing AFRP may withstand impact magnitudes beyond that expected for normal activities of daily living following IVD repair. Further studies are warranted to evaluate the endurance limit of these constructs following repeated impact loads. Moreover, considering the significant role that interlamellar collagen and elastin fibers play in AF mechanics, (Isaacs et al., 2014) and incorporation and crosslinking to- and/or within the AFRP should be explored.

The second significant finding of this study was that the AFRP was able to prevent native NP and AF tissue herniation upon the application of physiological loading. In order to restore mechanical function to the IVD following a discectomy, oAF repair strategies must prevent recurrent herniations of either native NP and AF tissue and/or implant surrogates to restore the mechanical interplay between the two regions. (Adam et al., 2015; Zahari et al., 2017) Previous studies demonstrate that resisting herniation upon application of axial compressive loads at or below physiological magnitudes remains a significant challenge for AF repair implants. (Cruz et al., 2018; Kang et al., 2017; Long et al., 2016, 2017) While some have demonstrated the ability of AF repair strategies to resist expulsion from focal AF defects, (Cruz et al., 2018; Likhitpanichkul et al., 2014) up to 30% of the native NP tissue was removed prior to AF repair and mechanical testing. As a result, the IVDs were likely depressurized, and thus, require the application of higher axial loads to produce tissue herniation/failure. Therefore, potentially inflating failure strength values. Alternatively, herein it has been demonstrated that the AFRP was able to resist the coordinate pressurization and radially-directed bulging of native NP and a full-thickness AF tissue plug under physiological loading. Pressurization of the IVD was confirmed by immediate ejection of the AF tissue plug upon application of minimal loads during pilot testing performed without an AFRP. Additionally, the coordinate interaction between NP and AF tissue was confirmed via video analysis, which demonstrated an outward bulge of the AFRP during axial compressive loading. No failure of the AFRPs was noted at the patch/suture interface (e.g. detachment). Additionally, no overt rupture of the patch material nor herniation/extrusion of the native NP or full-thickness AF tissue plug occurred during kinematic testing. Future studies should directly measure radial bulge magnitude and herniation potential during a combined compression-bending test protocol as is known to commonly cause herniation concomitant with measuring real-time intradiscal pressure.

As a result of its ability to effectively sequester the native NP and AF tissue within the IVD, the third significant finding from these studies was that the AFRP was able to assist in restoring FSU axial kinematic parameters that had been significantly altered due to injury. More specifically, annulotomy resulted in significant changes in the long-term creep response and creep displacement. This is generally in alignment with previous studies which demonstrated that injury to the NP, without disrupting the AF, caused significant changes in the early creep response with minimal changes occurring in the late response parameters, the latter of which the authors hypothesized were dictated by the AF. (Johannessen et al., 2006) Herein, we have confirmed this hypothesis and found that an isolated AF injury significantly changed the late creep response as compared to the intact condition. Conversely, only the early viscous damping coefficient was changed during annular injury which was likely the result of NP tissue migration in the defect path, not changes to the NP tissue properties. Furthermore, these parameters were restored following re-introduction of the AF tissue plug and sequestration with the AFRP. The significant decrease observed in creep displacement was also likely due to migration of the NP into the annular defect pathway. In axial compression, the NP is known to support low magnitude loads prior to engaging the AF, which supports high compressive loads via the generation of hoop stresses. (Johannessen et al., 2006) Since the NP migrated during compression, earlier transfer of compressive loads directly to the remaining AF are likely to occur, and thus, creep displacement would be reduced. Annular injury and NP migration also accounted for the decreased annular engagement and laxity of the FSU as indicated by the significant increase in neutral zone and ROM. Upon sequestering the NP tissue within its centralized location using the AF tissue plug and AFRP, the axial kinematic parameters that had significantly changed following annulotomy were now found not to be significantly different from the intact condition. Moreover, the repaired values were significantly different compared to the annulotomy group. Notably, however, torsional kinematic parameters were not restored following repair. This is likely the result of two phenomena; 1) a lack of tissue integration between the AF tissue plug and the adjacent AF, and 2) the inability to restore circumferential residual pre-strains found within the intact AF. (Mengoni et al., 2017; Michalek et al., 2012) The former could be addressed by using a composite repair strategy using a mimetic scaffold to plug the defect pathway that is adhered to adjacent AF tissue using durable adhesive hydrogels prior to securing the implants within the IVD using an AFRP.

Finally, compression to failure testing demonstrated that the AFRP was able to withstand stresses exceeding those reported within the human lumbar IVD during activities of daily living. (Wilke et al., 1999) Due to the longitudinal testing scheme used for this study, we did not assess the failure strength of intact FSUs, which could be considered a study limitation. However, others have demonstrated the compressive failure strength of intact FSUs to exceed 10MPa. (O’Connell et al., 2015)

Additional limitations were noted herein. Concerning kinematic testing, FSUs were not re-equilibrated through free swelling while submerged in a PBS bath between experimental group testing which has been shown to impact kinematic parameters (Bezci and O’Connell, 2018). Conversely, herein FSUs were recovered in an unloaded state for a period of eighteen-hours wrapped in gauze soaked in PBS/protease inhibitor, and thus, it was assumed that the IVDs could imbibe the solution. Additionally, all kinematic parameters were tracked longitudinally in the same IVDs via a repeated measures study design, and thus, any effect of rehydration was carried across all testing groups. An additional limitation with the kinematic testing was a lack in specifically tracking IVD height changes that may have occurred during the transition between the mechanical test frames that were used for performing creep testing (performed on a Bose Electroforce) and cyclic testing (performed on an Instron) in these studies. However, the same compressive pre-stress (−0.125MPa) applied at the end of the creep test was re-applied prior to the start of the cyclic testing which occurred within three (3) minutes of each other. This helped ensure that the IVD height was the same despite switching test frames. Furthermore, post-test evaluations of the displacement levels at the mean amplitude of −0.125 MPa following creep loading and at the beginning of the cyclic loading segments was not found to be significantly different (0.423±0.029 mm and 0.422±0.105 mm, respectively; p=0.9779). Lastly, the authors acknowledge that laboratory-based mechanical testing performed herein does not directly reflect the complex loading and kinematics which occur in vivo.

Conclusions

In summary, the inclusion of a GAG-based ILM significantly enhanced the radial impact burst strength of the AFRP compared to those without. Furthermore, the scaffolds exhibited the strength to prevent herniation of native (i.e. fully pressurized) NP and AF tissue while resisting detachment from adjacent oAF tissue. By being able to sequester native IVD tissue, the AFRPs were able to contribute to the restoration of IVD axial kinematic parameters that had changed due to isolated annular injury. Taken together, this data suggests that AFRPs are a viable option to prevent the herniation of NP and AF tissue and/or scaffolds used in the composite repair of IVDs to restore mechanical function following discectomy.

Highlights.

• Incorporation of an interlamellar matrix increases radial impact strength

• Focal defects require full-thickness and an outer closure system repair

• Repair patch prevents herniation of native tissues under physiological loading

Abbreviations

AFRP

Annulus Fibrosus Repair Patch

FSU

Functional Spinal Unit

IVD

Intervertebral Disc

IVDH

Intervertebral Disc Herniation

AF

Annulus Fibrosus

oAF

Outer AF

iAF

Inner AF

NP

Nucleus Pulposus

NPR

Nucleus Pulposus Replacement

GAG

Glycosaminoglycan

ILM

Interlamellar Matrix

IDP

Intradiscal Pressure

HA

Hyaluronic Acid