The Effect of Mechanical Stress on Ligamentum Flavum Hypertrophy; Utilization of Novel In Vitro Multi-torsional Stretch Loading Device

Objective We developed a novel multi-torsional mechanical stretch stress (MSS) loading device for ligamentum avum (LF) cells and evaluated its inuence on the development of ligamentum avum hypertrophy (LFH), a common cause of lumbar spinal canal stenosis. Materials and Methods Stretch strength of the device was optimized by applying 5% and 15% MSS loads for 24, 48, and 72 h. A cytotoxicity assay of human LF cells was performed and the results were compared to control (0%) MSS. Inammatory markers (interleukin [IL]-6, IL-8), vascular endothelial growth factor [VEGF], and extracellular matrix (ECM)-regulating cytokines (matrix metalloproteinase [MMP]-1, MMP-3 and MMP-9, and tissue inhibitor of metalloproteinase [TIMP]-1 and TIMP-2) were quantied via enzyme-linked immunosorbent assay. Using our multi-torsional MSS loading device, 5% MSS for 24 h was optimal for LF cells. Under this condition, the IL-6 and IL-8 levels, VEGF level, and MMR-1, MMR-3, and TIMP-2 were signicantly increased, compared to the control. Using the novel multi-torsional MSS loading device we conrmed that, mechanical stress enhances the production of inammatory cytokines and angiogenic factors, and altered the expression of ECM-regulating enzymes, possibly triggering LFH. This discovery enhances our understating of the effects of mechanical stress on LFH.


Introduction
Up to 70% of the general population experiences chronic lower back pain (LBP) once or more throughout their lifetime. 1 Among the many possible causes of chronic LBP, the prevalence of lumbar spinal canal stenosis (LSCS) in the elderly population is gradually increasing. 2 LSCS is also associated with lowerextremity radiculopathy and neurogenic claudication, which greatly affects the walking distance of the elderly. These clinical symptoms are associated with daily quality of life and therefore are of great interest to spinal physicians. The pathomechanism of LSCS is unclear, but facet joint enlargement, central intervertebral disc bulging, and ligamentum avum hypertrophy (LFH) are contributing factors. 3 Among them, LFH secondary to the aging process or mechanical stimulation induced by instability of the spinal segment are key. 4 Therefore, research on the physiologic basis of LFH has caught the attention of spinal specialists, who agree that in ammation, angiogenesis, and matrix regulation of ligamentum avum in uence the development of LFH. [5][6][7][8][9] Mechanical stress on the ligamentum avum is a major contributing factor to LFH. Hayashi et al. reported that mechanical stress concentration was directly linked to LFH in a rabbit model, 10,11 and Hur et al. emphasized the link between angiogenesis and mechanical stress-induced LFH. 7 Other studies have revealed an association between in ammation triggered by mechanical stress and LFH. 6-9 Nonetheless, it is doubtful whether these studies mimic in vivo mechanical stress. In this study, we developed a novel multi-torsional cell plate stretch device that mimics in vivo mechanical stress on ligamentum avum tissue. We evaluated the molecular biological responses related to in ammation, angiogenesis, and extracellular matrix (ECM) regulation of ligamentum avum cells to various stress loads to identify the stress load that best mimics LFH.

Optimizing MSS Load on LF Cells
Using dual-step motor generators controlled by motor drivers, multi-torsional MSS was successfully loaded on the assembled cell chambers (Video File Supplemental Data).
The tension-load was produced by optimizing the multi-torsional stretch strength and the cyclic load frequency, and 3D simulation was performed to visualize the expected load on the chambers (Fig. 4). Effect of MSS load on production of ECM-regulating factors MMP-1, MMP-3, MMP-9, TIMP-1, and TIMP-2 release levels from LF cells loaded with MSS at 5% for 24 h were measured to assess ECM remodeling. Following MSS loading, MMP-1, MMP-3, and MMP-9 release levels were 463.94 ± 53.08, 579.92 ± 90.43, and 25.77 ± 1.84 ng/mL, respectively. The TIMP-2 release level was 320.00 ± 16.34 ng/mL; TIMP-1 was undetectable. The MMP-1, MMP-3, and TIMP-2 release levels were signi cantly increased by MSS loading compared to the control group (330. 15  We previously reported that in ammation and subsequent angiogenesis are involved in the pathomechanism of LFH in vitro, indicative of close relationships among in ammation, angiogenesis, and LFH. 6 In a follow-up study of the association between in vitro and clinical data, we discovered links among mechanical stress, angiogenesis, and LFH. 7 However, these studies were limited in that mechanical stress was not loaded directly onto the LR cells. Instead, the effects of mechanical stress were evaluated indirectly based on radiological ndings. In this study, we developed a novel mechanical stress loading device with multidirectional torsion that mimics the mechanical load on LF tissue in vivo. Rather than inducing in ammation by transforming growth factor-β1 (TGF-β1) or interleukin-1β (IL-1β), we used mechanical stress on the LF and believe it re ects the effects of mechanical stress on LFH.
Our results demonstrated that multi-torsional MSS load for 24 h under 5% stretch force stimulation resulted in an increase in IL-6 and VEGF levels. IL-6 activates neutrophils, whose adhesion and brosis are promoted by increased expression of ECM-regulating molecules or cytokines. 17  that MMP-9 expression is higher in LFH tissue. 19 TIMPs also regulate ECM homeostasis, and TIMP-1 and TIMP-2 play key roles in brosis in various cell types by increasing proliferation. Park et al. hypothesized that TIMP-1 and TIMP-2 in uence LFH by increasing ECM density and promoting hypertrophy by suppressing MMP activities. 4 This hypothesis was con rmed by the signi cant association between elevated TIMP-1 and TIMP-2 expression in LF broblasts and spinal stenosis, a reproducible nding of several different experiments of various methods. 27 This is compatible with our TIMP-1 and TIMP-2 expression data.
Mechanical stress is a key factor in LFH, as con rmed by in vitro 28-32 , in vivo 10,11 , and clinical studies 7 .

Chao et al. developed an in vitro method of loading stress on LF cells by centrifuging them in a horizontal
microplate rotor. 32  to pull a exible cell culture plate from the center. 30 It is meaningful that centrifugal and cyclic onedimensional mechanical forces on LF broblasts affected the mechanostress pathway. However, because one-and two-dimensional forces are unlike that on LF tissue in vivo, the accuracy of the model is unknown. Therefore, it is signi cant that we developed a reproducible repetitive mechanical stress loading device that recapitulates the mechanical stress on LF cells. The device will be used to provide insight into the role of direct mechanical stress on LFH in vitro and the cells' fate after mechanical stress loading.

Ethic declaration
This study was reviewed and approved by the local ethics committee (Research Ethics Committee of Korea University Guro Hospital: approval number K2017-0991) and has been performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments.
Informed consent was obtained from all participants.

Human LF Cell Isolation and Culture
This study was approved by the Institutional Review Board (IRB) of our institute. Human LF tissues were collected during surgeries on the lumbar spine for herniated nucleus pulposus, following the regulations of the IRB. LF cells were isolated from the tissues of ve patients of normal LF thickness. LF tissues harvested in the operating room were placed in sterile Ham's F-12 medium (Gibco-BRL, Grand Island, NY) containing 1% penicillin/streptomycin (P/S; Gibco-BRL) and 5% fetal bovine serum (FBS; Gibco-BRL). After a phosphate-buffered saline (PBS; Welgene, Gyeongsan-si, Gyeongsangbuk-do, Korea) wash, tissues were minced and digested for 1 h at 37°C in Dulbecco's modi ed Eagle's medium (DMEM; Welgene, Gyeongsan-si, Gyeongsangbuk-do, Korea) with 0.2% pronase (Calbiochem, La Jolla, CA). Next, LF tissues were incubated overnight at the same temperature in 0.025% collagenase I (Roche Diagnostics, Mannheim, Germany). LF cells were ltered through a sterile nylon-mesh cell strainer (pore size, 70 µm), centrifuged, and the pellets were resuspended and cultured in DMEM containing 10% FBS and 1% P/S in a humidi ed atmosphere of 5% CO 2 at 37°C. LF cultures were continued until reaching full con uence.
The cells were trypsinized and replated for subculture. Subsequent experiments were conducted using these second-passage LF cells.

Design and Implementation of the Novel Mechanical Stretch Stress Loading System
We fabricated a multiple-multidirectional mechanical stretch stress (MSS) loading chamber system capable of incubating dishes containing LF cells. The multi-torsional cell plate stretch device comprises a roo ess metal frame containing xation panels, twisting parts, culture chambers, and a controller.
Multiple chambers are seated parallel on the xation panel facing upwards (Fig. 1). The sides of the chamber are xated to two separate and parallel-oriented xation panels, which pulls the chamber by moving in the opposite direction. In addition, the xation panels are coupled to the twisting part to produce torsion stress on multiple chambers. The parallel chambers are aligned and stretched in the same direction and with identical power simultaneously. Each chamber is made of exible polydimethylsiloxane (PDMS) by photolithography, that can contain cell cultures and stretch or twist. An optically transparent, ultrathin (100 µm) membrane was applied to the well bottom to render the stretch chambers compatible with optical and uorescence microscopy (Fig. 2). The MSS force developed by two step motor generators were controlled by Arduino Uno and L293D motor drivers, regulating the strength of the stretch and torsional stress. The optimal cyclic directions and loading were established after multiple virtual simulations. A 4-degree tilt away from the panel provides 2 mm stretch and 3 mm sliding of each panel beneath the chambers, resulting a in 10 degree of rotation tilt of the chamber corners and torsional stretch on the PDMS chambers (Fig. 3). The PDMS chambers are designed to be assembled on the MSS device after cell attachment has been con rmed. In addition, to determine the expected stretch load force on the internal surface of the chamber, a three-dimensional (3D) simulation program (Inventor, Autodesk Inc, CA) was used. The torsional stress loaded on the external chamber surface was analyzed and presented as stress-strain ratios.
Mechanical Stretch Stress Loading on LF cells LF cells were plated on the PDMS chamber at a density of 1.0 ⋅ 10 4 /mL. After 24 h of incubation, cell attachment to the cell chamber wall was veri ed, and the cells were subjected to MSS. Multi-torsional MSSs of 0% (no stretch-control), 5%, and 15% of maximal stretch load were applied to multiple cell plate chambers simultaneously.

Statistical Analysis
Data are means ± standard deviations (SDs) for individual experiments using independent cell cultures. P values were calculated using Student's t-test or the Mann-Whitney U test, as appropriate according to sample size and distribution normality. P-values < 0.05 were considered to indicate statistical signi cance.

Declarations
Author contributions KWK, HCH equally assisted with study design, data collection, data interpretation, and drafting of the manuscript. CH, BSM, LJW, PYK, MHJ assisted with study design, data collection, data interpretation, KJH is the corresponding author and assisted with study design, data collection, data interpretation and drafting of the manuscript as well.

Figure 1
An open metal frame with multiple chambers seated parallel to the xation panels.  Three-dimensional simulation of the expected load on the chambers. The torsional stress loaded on the external chamber surface is shown as the stress-strain ratio.

Figure 5
Cytotoxicity assay of ligamentum avum cells. MSS loading at 5% stretch for 24 h on ligamentum avum cells did not signi cantly affect the LDH release level. Stronger stretch (15%) and stretching for 48 h resulted in signi cantly increased LDH levels, indicating a cytotoxic effect.