Preclinical in vivo animal models of intervertebral disc degeneration. Part 1: A systematic review

Abstract

Intervertebral disc degeneration (IVDD), a widely recognized cause of lower back pain, is the leading cause of disability worldwide. A myriad of preclinical in vivo animal models of IVDD have been described in the literature. There is a need for critical evaluation of these models to better inform researchers and clinicians to optimize study design and ultimately, enhance experimental outcomes. The purpose of this study was to conduct an extensive systematic literature review to report the variability of animal species, IVDD induction method, and experimental timepoints and endpoints used in in vivo IVDD preclinical research. A systematic literature review of peer‐reviewed manuscripts featured on PubMed and EMBASE databases was conducted in accordance with PRISMA guidelines. Studies were included if they reported an in vivo animal model of IVDD and included details of the species used, how disc degeneration was induced, and the experimental endpoints used for analysis. Two‐hundred and fifty‐nine (259) studies were reviewed. The most common species, IVDD induction method and experimental endpoint used was rodents(140/259, 54.05%), surgery (168/259, 64.86%) and histology (217/259, 83.78%), respectively. Experimental timepoint varied greatly between studies, ranging from 1 week (dog and rodent models), to >104 weeks in dog, horse, monkey, rabbit, and sheep models. The two most common timepoints used across all species were 4 weeks (49 manuscripts) and 12 weeks (44 manuscripts). A comprehensive discussion of the species, methods of IVDD induction and experimental endpoints is presented. There was great variability across all categories: animal species, method of IVDD induction, timepoints and experimental endpoints. While no animal model can replicate the human scenario, the most appropriate model should be selected in line with the study objectives to optimize experimental design, outcomes and improve comparisons between studies.

1. INTRODUCTION

The intervertebral disc (IVD) is a complex fibrocartilaginous tissue that unites adjacent vertebrae along the spine to facilitate axial load‐bearing and spinal motion. ¹ Intervertebral disc degeneration (IVDD), a widely recognized cause of lower back pain (LBP), is a common musculoskeletal disorder with a range of impacts and the potential to cripple those affected both functionally and financially. The prevalence of IVDD correlates with age, occurring in over 70% of people under 50 years of age and in over 90% of those over 50. ² LBP restricts millions of people each year with an estimated socioeconomic burden of $90 billion annually in the United States alone. ³ LBP is the leading medical condition causing disability worldwide ⁴ and with an aging population, further research to better understand and accurately model IVDD is essential for exploring therapeutic treatment of this condition.

In the absence of spontaneous reproducible IVDD models, preclinical in vivo animal models have been developed and used over many years to further our understanding of the pathophysiology of IVDD with the aim to ultimately to develop and evaluate treatment strategies. ⁵ Common preclinical IVDD animal models include the use of rodent, rabbit, canine, ovine, porcine, primate and caprine species. ⁶ In these studies, there is great variability in the selection of animals used, method of inducing disc degeneration and measurable experimental endpoints being assessed. The ideal animal model should represent similar morphological and biomechanical properties to that of the human IVD. ⁷ However, there are many differences between animal species and humans including but not limited to; cell population, tissue composition, disc anatomy, spinal anatomy, growth and development and functional biomechanics. ⁸ A diverse array of IVDD methods have since been explored in varying animals with researchers employing models that feature surgical or chemical disc injury, non‐invasive mechanical stress, genetic modifications and natural occurrence. ⁵ Additionally, there is significant variation in the diagnostic endpoints employed by preclinical studies to evaluate the efficacy of their preclinical animal models for IVDD. Endpoints utilized vary from imaging modalities such as radiography, computed tomography (CT) and magnetic resonance imaging (MRI) to histopathology, biochemical assays, immunohistochemistry, mechanical testing, and a variety of other measurable outcomes. ⁹ , ¹⁰

While no animal model perfectly parallels human IVDD, careful selection of a reproducible and valid animal model is required for applicable results that can narrow the gap between the research and clinical settings. ⁵ , ⁶ There is a need to critically evaluate the available preclinical in vivo models of IVDD in the literature to better inform researchers and clinicians in the field and to encourage the practice of sound scientific rationale for the conduction of these models. A thorough search of the literature did not reveal any in vivo animal model reviews that appropriately summarized the animal species used, method of induction of IVDD, timepoints and experimental endpoint measures. The aim of this study was to report the variability of animal species, method of achieving IVDD, timepoints and experimental measures employed in IVDD preclinical research via systematic review. We hope that this review serves as a reference point for researchers attempting to standardize preclinical animal IVDD models and to achieve future relatability across IVDD studies.

2. MATERIALS AND METHODS

2.1. Methodology

A systematic review that adhered to the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines was performed on September 30, 2022. ¹¹ , ¹² Details of the protocol for this systematic review were registered on PROSPERO and can be accessed at https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=282284.

Ethical approval was not required for this study.

2.2. Selection criteria

This study analyzed animal models of IVDD. Studies were included in the systematic review if they described a novel animal model of disc degeneration, an established model to evaluate therapeutic repair or proposed modifications to existing models. Novel methods were identified by the authors during the screening process. Modifications were noted if the model included the use of the same animal species for a different degeneration method, or if there was significant variation in the specific degeneration technique or animal characteristics such as age. The publication language was restricted to English. Duplicates were removed and after screening the titles and abstracts for relevance to the study question, full‐text original articles were sought for retrieval. The studies required the inclusion of a description about the animal species, method of inducing disc degeneration and measured outcomes. The reported methods for modeling disc degeneration included in this systematic review are as follows: surgical, chemical, genetic, and bacterial interventions, spontaneous degeneration from the natural aging process and non‐invasive induction methods including behavioral, postural, and dietary modifications. All the retrieved publications were independently reviewed by two investigators against the inclusion criteria to determine eligibility. Discrepancies were resolved by a third investigator.

2.3. Search strategy

Records published through to September 2022 were searched on the electronic databases PubMed and EMBASE, using the following search terms: (animal model OR model) AND (spondylosis OR lumbar disc degeneration OR intervertebral disc hernia OR spinal degeneration OR disc degeneration OR intervertebral disc disease OR degenerative disc disease OR intervertebral disc degeneration) NOT (mitigate OR protect OR attenuate OR alleviate OR macula). To identify any sources missed during the literature search, the reference list of the included articles was also screened; however, no additional articles were obtained.

2.4. Data collection

Two investigators independently extracted data from the included papers using a standardized form. The species and age of the animals, timepoints used in the study, the degeneration stimulus, and techniques used to measure outcomes were collected, and the results were recorded descriptively.

3. RESULTS

3.1. Study selection

The literature search yielded 1849 results from EMBASE and 1619 from PubMed, representing 3025 unique reports for screening. Following screening according to inclusion criteria, 259 studies were identified for inclusion in the systematic review (Figure 1).

FIGURE 1.

Flowchart of the results based on selection and inclusion of manuscripts through database searching and record screening

The publication dates of the included articles span 40 years, ranging from 1983 to 2022 inclusively. A total of 51.74% (134/259) of the articles were published within the last 8 years (Figure 2).

FIGURE 2.

Date of publication by year of the articles reviewed in this systematic review

3.2. Type of animal model and species used

Two‐hundred and fifty‐nine (259) preclinical in vivo animal models of IVDD were reviewed. The breakdown of models by animal species is shown in Figure 3. Rodents (mice or rats) were the most common animal species chosen (140/259, 54.05%). The second most common animal species chosen was rabbits (56/259, 21.62%). The majority of studies reported the breed of animal species (237/259, 91.51%). Age of the animals used was reported in 216/259 (83.40%) of studies. For each species, the age range of animals enrolled in any animal model design, excluding manuscripts utilizing spontaneous degeneration, were as follows: dog—6–15 months ¹³ , ¹⁴ ; goat—8 weeks to 4 years ¹⁵ , ¹⁶ ; monkey—5.5–9 years ¹⁷ , ¹⁸ ; pig—4 weeks to 2 years ¹⁹ , ²⁰ ; rabbit—1.5–19 months ²¹ , ²² ; rodent—2 weeks to 13.5 months ²³ , ²⁴ ; sheep—11 months to 5 years. ²⁵ , ²⁶

FIGURE 3.

Species employed expressed as a percentage of the total number of research papers incorporated in this systematic review. *Includes a single manuscript each for the species of chicken, guinea pig and horse

3.3. Method of IVDD induction

The methods to induce IVDD across all studies are shown in Figure 4A. Surgical induced injury was the most common (168/259, 64.86%), followed by chemical injury (46/259, 17.76%) genetically modified (28/259, 10.81%) and spontaneous (28/259, 10.81%) models. For chemically induced degeneration, the two most common substances used were enzymes, via intradiscal, subchondral or systemic injection, ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² and poly (methyl methacrylate) (PMMA), via intradiscal PMMA particle injection or into the vertebral body parallel to the adjacent vertebral endplate ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ (Figure 4B). Of the 28 (10.81%) papers that reported genetic modification to induce IVDD, 20 different genes were targeted for either gene knock‐out, ²³ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ , ⁵⁴ , ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² , ⁶³ , ⁶⁴ , ⁶⁵ , ⁶⁶ , ⁶⁷ , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ gene knock‐in ⁷² or gene over expression. ⁷³ Another category of achieving IVDD was classified as spontaneous, totalling 10.81% (28/259), which encompassed all studies that obtained animal specimens with naturally occurring IVDD. Of these spontaneously induced IVDD studies, there was great variation in animal species selection, with the following animals incorporated as the subjects of these studies: rat, ⁷⁴ , ⁷⁵ , ⁷⁶ , ⁷⁷ , ⁷⁸ , ⁷⁹ , ⁸⁰ mouse, ⁴⁸ , ⁸¹ , ⁸² , ⁸³ rabbit, ⁸⁴ , ⁸⁵ , ⁸⁶ dog, ⁸⁷ , ⁸⁸ , ⁸⁹ , ⁹⁰ sheep, ⁹¹ , ⁹² pig, ⁹³ horse, ⁹⁴ and monkey. ⁹⁵ , ⁹⁶ , ⁹⁷ Notably, needle puncture was the dominant surgical method, being used in 24.71% (64/259) of the studies, with a large number of other surgical methods also being used (Figure 4C).

FIGURE 4.

Summary of methods of degeneration utilized to simulate IVDD expressed as a percentage of the total number of research papers incorporated in this systematic view. (A) general categories of injury methods; (B) specific methods of chemically induced injury to simulate IVDD; (C) specific types of surgically induced injury to simulate IVDD. *Includes hindlimb unloading (vertical vibration plate or tail‐suspension), axial loading (bi‐pedal posture with water limiting space, hot plate, animal centrifuge device or custom enclosure with a weighted collar) and dietary modifications. **Includes Ilizarov‐type implants, Kirschner wires for the application of a custom‐made external loading device, scoliosis implants (staples and bone anchor ligament tethers), titanium‐screw‐based chronic compression device, intervertebral distraction screw, pedicle screw with stainless steel rod or cable and stainless‐steel wire. ^†Includes exposure to nicotine or tobacco, injection of diethylstilbesterol, tamoxifen, fibronectin, methylprednisolone, streptozotocin, botulinum toxin, pingyangmycin, lidocaine, cytokines, camptothecin, bleomycin, complete Freund's adjuvant, leptin, ethanol or disc cells transfected with adenovirus. ^‡Includes annulus puncture with a biopsy gun, Kirschner wire, tenotomy knife or trephine, articular processes cut and sutured, annulectomy, forelimb amputation, brachial plexus rhizotomy, tail amputation or circumcision and sutured for compression, soft tissue surgical exposure alone, laminotomy, implanted bullet fragments, neonatal pinealectomy and manual torsion applied intraoperatively

3.4. Experimental endpoints

Experimental endpoints used in the reviewed studies are summarized in Figure 5A. Histology was the most common endpoint, being used in 83.78% (217/259) of studies. The second most used endpoint was immunohistochemistry at 45.56% (118/259). The most common diagnostic imaging modalities included T1‐ and T2‐weighted MRI (112/259, 43.24%), radiography (103/259, 39.77%), micro‐CT (37/259, 14.29%) and CT (9/259, 3.47%). Additional imaging modalities included micro‐MRI, bioluminescent imaging, second harmonic generation imaging, angiography, discography, DEXA (dual‐energy x‐ray absorptiometry) bone density scans, molybdenum target x‐ray spectra and infrared imaging. These modalities were used in a combined total of 5.02% (13/259) of the reviewed publications. There was significant variation in the type of histological stain employed by the varying methodologies, as shown in Figure 5B. Hematoxylin and eosin was the most common histological stain used; 62.55% (162/259) of studies. Biochemical assays were used in 12.74% (33/259) of the studies and are summarized in Figure 5C. Time points of when each experimental methodology was used for outcome evaluation varied greatly depending on the animal species and method of injury employed to attain IVDD (Appendix A).

FIGURE 5.

Experimental endpoints reported expressed as a percentage of the total number of research papers incorporated in this systematic review. (A) general categories of outcomes measured; (B) specific types of histology stains used for histological analysis; (C) specific types of biochemical assays used for biochemical analysis. *Includes Hoechst 33258 and PicoGreen assays for DNA content. **Includes the Blyscan assay and assays for uronate and hexosamine content to determine proteoglycan concentration. ^†Includes micro‐MRI, flow cytometry, bioluminescent imaging, fluorescent microscopy, vertebral endplate permeability, atomic force microscopy, second harmonic generation imaging, cell counting kit‐8, computer generated finite element model, MR microscopy, FTIR spectroscopy, in situ hybridisation, gene transcription analysis, electrophoresis, digital subtraction angiography, vertebral transcorporeal discography, DEXA scan, infrared imaging, blood test, energy dispersive x‐ray spectroscopy, x‐ray diffraction analysis, nerve conduction velocity, mitochondrial function, manual palpation, gel chromatography, molybdenum target x‐ray spectra, and microarray. ^‡Includes Alizarin red, tartrazine, trypan blue, methylene blue‐Azure II, methyl green, Heidenhain trichrome, von kossa, giemsa, luxol fast blue, tetrachrome, and tartrate‐resistant acid phosphatase (TRAP) staining. ^ Includes MTT assay, glucose‐oxidase assay, RIPA assay, gelatin zymography assay, lysosome activity assay, caspase 3/7 assay, and EdU‐chase assay

3.5. Experimental timepoints

A breakdown of the selected timepoints in the reviewed preclinical in vivo models for all species and method of IVDD induction is summarized in Table 1. The time that degenerative changes were evaluated post induction in the reviewed animal models varied greatly between species used and ranged from 1 week in the dog and rodent, up to greater than 104 weeks (or 2 years) in the dog, horse, monkey, rabbit, rodent, and sheep. Timepoints for each individual species were considered common if five or more manuscripts reported identifying IVDD. The most commonly selected timepoint across all species for all preclinical IVDD models was 4 weeks (49 manuscripts). Common timepoints by species were the following: goat—12 weeks (8 manuscripts); monkey—greater than 104 weeks (7 manuscripts); pig—12 weeks (7 manuscripts); rabbit—2 (7 manuscripts), 4 (17), 12 (14) and 24 (5) weeks; rodent—1 (18 manuscripts), 2 (24), 3 (5), 4 (24), 8 (19), 12 (16), 16 (5), 20 (5) and 24 (7) weeks; sheep—24 weeks (5 manuscripts). The number of manuscripts per experimental timepoint were also charted in reference to the specific method of degeneration utilized for the two most used species being rodents (Table 2) and rabbits (Table 3). For in vivo models where IVDD was surgically induced in rodents, the most common timepoints were 1 (16 manuscripts), 2 (20), 4 (15), 8 (8) and 12 (11) weeks (Table 2). The most common experimental timepoints of identifying degenerative changes in other rodent models were the following: bacterial—1 week (2 manuscripts); chemical—4 weeks (4 manuscripts); genetic—8 weeks (5 manuscripts); non‐invasive—4 weeks (4 manuscripts); spontaneous—72 weeks (3 manuscripts). For in vivo models where IVDD was surgically induced in rabbits, the most common timepoints were 2 (6 manuscripts), 4 (15) and 12 (11) weeks (Table 3).

TABLE 1.

Number of manuscripts per experimental timepoint (weeks) for all in vivo preclinical intervertebral disc degeneration (IVDD) models in all species

Weeks	No. of manuscripts
	Chicken	Dog	Goat	Guinea Pig	Horse	Monkey	Pig	Rabbit	Rodent	Sheep
1	‐	2	‐	‐	‐	‐	‐	‐	18	‐
2	‐	‐	‐	‐	‐	‐	‐	7	24	‐
3	‐	‐	‐	‐	‐	‐	2	1	5	‐
4	‐	1	‐	‐	‐	1	2	17	24	4
5	‐	‐	‐	‐	‐	‐	1	1	‐	‐
6	‐	‐	‐	‐	‐	‐	‐	3	5	3
7	‐	‐	‐	‐	‐	‐	‐	‐	1	‐
8	1	‐	1	1	‐	‐	4	4	19	1
9	‐	‐	‐	‐	‐	‐	‐	2	2	‐
10	‐	‐	‐	‐	‐	‐	‐	‐	3	‐
12	‐	1	8	‐	‐	‐	7	14	16	‐
15	‐	‐	‐	‐	‐	‐	‐	2	1	‐
16	‐	1	‐	‐	‐	‐	‐	‐	5	1
17	‐	‐	‐	‐	‐	‐	‐	‐	‐	1
18	‐	1	‐	‐	‐	‐	‐	‐	1	‐
20	‐	‐	‐	‐	‐	‐	‐	‐	5	‐
21	‐	‐	‐	‐	‐	1	‐	‐	‐	‐
24	‐	‐	3	‐	‐	4	2	5	7	5
28	‐	‐	‐	‐	‐	‐	1	1	‐	‐
32	‐	‐	‐	‐	‐	‐	1	‐	‐	‐
36	‐	‐	‐	‐	‐	‐	‐	‐	2	‐
42	‐	‐	‐	‐	‐	‐	‐	‐	‐	1
44	‐	‐	‐	‐	‐	‐	‐	‐	2	‐
46	‐	‐	‐	‐	‐	‐	‐	‐	‐	2
48	‐	‐	2	‐	‐	‐	‐	2	3	‐
52	‐	1	‐	‐	‐	‐	‐	‐	1	1
56	‐	‐	‐	‐	‐	‐	‐	‐	1	‐
60	‐	‐	‐	‐	‐	‐	‐	1	‐	‐
68	‐	1	‐	‐	‐	‐	‐	‐	‐	‐
72	‐	‐	‐	‐	‐	‐	‐	‐	4	‐
74	‐	‐	‐	‐	‐	‐	‐	‐	1	‐
76	‐	‐	‐	‐	‐	‐	‐	‐	1	‐
78	‐	‐	‐	‐	‐	‐	‐	‐	1	‐
88	‐	‐	1	‐	‐	‐	‐	‐	‐	‐
96	‐	‐	‐	‐	‐	‐	‐	2	1	‐
> 104 weeks (2 years)	‐	1	‐	‐	1	7	‐	1	1	1

Note: Yellow highlight (<8 weeks) and green highlight (>8 weeks).

TABLE 2.

Number of manuscripts per experimental timepoint (weeks) and method of degeneration for all in vivo pre‐clinical intervertebral disc degeneration (IVDD) models in rodents (mice and rats)

Weeks	No. of manuscripts
	Bacterial	Chemical	Genetic	Non‐invasive a	Spontaneous	Surgical
1	2	‐	‐	‐	‐	16
2	‐	2	1	‐	1	20
3	‐	1	‐	‐	‐	4
4	‐	4	1	4	‐	15
5	‐	‐	‐	‐	‐	‐
6	‐	‐	1	‐	‐	4
7	‐	1	‐	‐	‐	‐
8	‐	3	5	1	2	8
9	‐	1	‐	‐	‐	1
10	‐	‐	‐	2	‐	1
12	1	1	2	‐	1	11
15	‐	‐	1	‐	‐	‐
16	‐	1	1	‐	1	2
17	‐	‐	‐	‐	‐	‐
18	‐	‐	1	‐	‐	‐
20	‐	‐	3	‐	‐	2
21	‐	‐	‐	‐	‐	‐
24	‐	1	3	‐	‐	3
28	‐	‐	‐	‐	‐	‐
36	‐	1	‐	‐	‐	1
44	‐	‐	‐	‐	‐	2
48	‐	‐	2	‐	1	‐
52	‐	‐	1	‐	‐	‐
56	‐	‐	‐	‐	1	‐
60	‐	‐	‐	‐	‐	‐
72	‐	‐	‐	1	3	‐
74	‐	‐	‐	‐	1	‐
76	‐	‐	1	‐	‐	‐
78	‐	‐	1	‐	‐	‐
96	‐	‐	‐	‐	1	‐
144	‐	‐	‐	‐	1	‐

Note: Yellow highlight (<8 weeks) and green highlight (>8 weeks).

^a Includes hindlimb unloading (vertical vibration plate and/or tail‐suspension), axial loading (bipedal posture with a water limiting space or hot plate) and dietary modifications.

TABLE 3.

Number of manuscripts per experimental timepoint (weeks) and method of degeneration for all in vivo preclinical intervertebral disc degeneration (IVDD) models in rabbits

Weeks	No. of manuscripts
	Bacterial	Chemical	Genetic	Non‐invasive a	Spontaneous	Surgical
1	‐	‐	‐	‐	‐	‐
2	1	‐	‐	‐	‐	6
3	‐	‐	‐	‐	‐	1
4	‐	2	‐	‐	‐	15
5	‐	1	‐	‐	‐	‐
6	‐	1	‐	‐	‐	2
7	‐	‐	‐	‐	‐	‐
8	1	2	‐	1	‐	‐
9	‐	‐	‐	‐	‐	2
10	‐	‐	‐	‐	‐	‐
12	‐	3	‐	‐	‐	11
15	‐	1	‐	‐	‐	1
16	‐	‐	‐	‐	‐	‐
17	‐	‐	‐	‐	‐	‐
18	‐	‐	‐	‐	‐	‐
20	‐	‐	‐	‐	‐	‐
21	‐	‐	‐	‐	‐	‐
24	‐	2	‐	1	‐	2
28	‐	1	‐	‐	‐	‐
36	‐	‐	‐	‐	‐	‐
44	‐	‐	‐	‐	‐	‐
48	‐	‐	‐	‐	1	1
52	‐	‐	‐	‐	‐	‐
56	‐	‐	‐	‐	‐	‐
60	‐	‐	‐	‐	1	‐
72	‐	‐	‐	‐	‐	‐
74	‐	‐	‐	‐	‐	‐
76	‐	‐	‐	‐	‐	‐
78	‐	‐	‐	‐	‐	‐
96	‐	‐	2	‐	‐	‐
144	‐	‐	‐	‐	1	‐

Note: Yellow highlight (<8 weeks) and green highlight (>8 weeks).

^a Includes hindlimb unloading (modified tail‐suspension) and axial loading (custom enclosure with a weighted collar).

4. DISCUSSION

The objective of this study was to conduct a systematic review of available literature of preclinical IVDD models to summarize the variety of species, method of degeneration and experimental endpoints chosen. Additionally, we aimed to provide recommendations for the appropriate choice of animal species and degeneration method to guide future research using preclinical animal models. This is the first systematic review that has analyzed the in vivo animal models of IVDD employed in preclinical spine research. We hope that this review will be a valuable resource for researchers and clinicians developing animal models to investigate IVDD and advance our current understanding in the spine field.

4.1. Types of animal models

4.1.1. Comparisons between animals and humans

The IVD is a fibrocartilaginous tissue with the purpose of connecting adjacent vertebrae, absorbing axial loads and permitting spinal mobility. ⁶⁴ Each IVD consists of three tissues: an inner gelatinous nucleus pulposus (NP), concentric lamellae forming an outer annulus fibrosus (AF) and hyaline cartilage endplates (CEPs) that are responsible for 80% of nutritional supply to the IVD by enabling metabolic diffusion to occur. ⁹⁸

The anatomical structure and function of animal IVDs is similar to humans, with some species more so than others. Some important differences include variation in geometrical dimensions, persistent growth plates, retained or altered notochordal cells and variations in extracellular matrix (ECM) glycosaminoglycan content, water content and collagen content. ⁸ , ⁸⁸ , ⁹⁹ , ¹⁰⁰ , ¹⁰¹ , ¹⁰² , ¹⁰³ , ¹⁰⁴ , ¹⁰⁵

4.1.2. Variety of IVDD models

In vivo animal models have provided essential tools to further our understanding of the pathogenesis of IVDD and continue to do so given limited access to human tissue or human model systems. Various animal models have been used to study disc degeneration, with different animals utilized to model different pathological pathways of IVDD. Specific animal models may be chosen for different reasons based on the study design, research question, surgical treatment being investigated, or other therapeutic screening methods being evaluated. An extensive list summarizing the key categories of these animal model studies is shown in Appendix A. These studies aimed to establish evidence of disc degeneration in various species, and subsequently identify similar features between their degenerative processes and that of humans on morphological, histochemical, and biomechanical levels.

The mammalian species involved in the models incorporated in this review include the mouse, rat, rabbit, pig, sheep, dog, goat, monkey, guinea pig, and horse. One paper also investigated the effect of surgical pinealectomy on IVDs in chickens. ¹⁰⁶ From a general standpoint, rabbits and rodents may not be suitable for modeling IVDD in humans purely based on their size limitations. However, many aspects of the IVD are size‐dependant and after scaling specific parameters to achieve clinical relevance, the use of these more accessible species is advantageous for further research in the spine field. ⁸ Other important factors to consider for interspecies differences include biochemical, morphological, and cellular properties of the IVD, along with segmental range of motion and biomechanical loading. No single animal model can replicate the entirety of these factors in an identical fashion to humans; however, provided the limitations of a selected animal species are recognized, the results can be reliably interpreted.

4.2. Considerations with animal models

Animal models are widely used to study disc degeneration based on accessibility and practicality to conduct in vivo experiments. The cost, availability, age, size, biomechanics, and ethical concerns of choosing or designing an appropriate preclinical model are fundamental considerations for researchers using animal subjects. As every animal model is a representation of a human IVD and pathologic process in some way, it is important to consider the limitations associated with each to achieve the most appropriate selection.

4.2.1. Cost and availability

The time frame needed for investigation, the availability, type, strain, and size of animal, along with associated housing and care requirements, are all factors affecting the cost‐effectiveness of a study. ¹⁰⁷ Where a larger animal may be desired, husbandry may be prohibitive, particularly if a high number of animals are required for a well powered sample study. Availability is also important, making small animals such as rodents an attractive option due to ease of breeding and thus their improved attainability, study reproducibility, relatively quick maturation and aging, and lower maintenance costs. Such practicality is considered within the literature where sample sizes are noticeably smaller for models utilizing larger animals, as in the Sasaki et al. study having 11 sheep versus various rodent models having closer to 20 or 30 in a sample population. ¹⁰⁸ Therefore, funding considerations can influence the choice of model to be utilized and becomes an important aspect when planning a study that features a preclinical animal model.

4.2.2. Age

The strong association between human IVDD and increasing age means the age of the animal relative to the human is of significance when selecting a model. As aforementioned, the point in an animal's life cycle influences the disc's anatomical components, particularly with regards to presence of notochord cells. As notochord cells persist through life in certain species and are thought to have a progenitor and/or protective mechanism against degeneration, ¹⁰⁹ the age at which each animal reaches skeletal maturity should thus be taken into account. From reviewing the literature in Appendix A, there was a lack of consistency between studies for age groups used intra‐species‐wise. The age variation of the two most common species of animals enrolled in IVDD studies, excluding spontaneous degeneration models, were 1.5 to 19 months in the rabbit ²¹ , ²² and 2 weeks to 13.5 months in the rodent. ²³ , ²⁴ In multiple studies, age was not reported and instead replaced with weight ranges of the animals used. Most likely this is due to ambiguity in the literature in terms of any standardized, recommended ages for various animal species across the model spectrum. Skeletal maturity is reported for rats at 2 months, rabbits between 4 and 6 months, pigs between 18 and 24 months, and cows at 24–30 months. ¹⁰¹ , ¹¹⁰ Nonhuman primates are the only species appearing to have clear reports of aging at a rate 3.5 times that of humans. ⁹⁵ , ¹¹¹ The authors are in agreement with Gruber (2009) in saying that young adult, skeletally mature animals are preferred for any model of IVDD by intervention to minimize the occurrence of other confounding physiological problems associated with aging. ¹¹² In those studies that require animals to be skeletally mature due to regulatory standards, this should be documented by the investigators via screening of growth plates such as in rabbit species. ¹¹³ Failure to do so may mean there is variation in the included animals that could influence the results of the study.

It is important to note that the timepoint of the IVDD model chosen should be done so with an understanding of the lifespan of the species chosen. For example, a 24‐week timepoint for a rodent model is not likely to reflect the same hallmarks of IVDD age‐related changes that a 24‐week timepoint would for a sheep model. This review identified 5, 7 and 5 manuscripts describing rabbit, rodent and sheep IVDD models respectively, with a 24‐week timepoint for IVDD. It is outside the scope of this review as to the intricacies of these manuscripts and to identify how a 24‐week timepoint would vary for each species and each outcome variable. But rather, it is worthwhile that the research be aware of this and select the most appropriate animal species and timepoint for the research question and aims of the proposed project.

Histological and biochemical changes have been shown to occur naturally during the aging process and therefore, an ideal model of IVDD would also occur with no experimental manipulation. Spontaneously occurring degeneration has been described in the Sand rat, ⁷⁷ mice, ¹¹⁴ Chinese hamsters, ¹¹⁵ nonhuman primates ⁹⁷ , ¹¹¹ and chondrodystrophic dog breeds. ⁹ Limitations of spontaneous IVDD models include difficulties in sourcing animals of a consistent age, the time and associated costs in housing these animals to the age at which degeneration occurs and the variation in pathology between animals given the spectrum of degenerative disease observed. While the genetic background of chondrodystrophic dog breeds has been well characterized (BANNASCH et al.) and research bred beagles can be readily purchased at specific ages from vendors, variability still arises within the time of onset and course of the degeneration pathway. The limitations described may influence the reliability of experimental outcomes and therefore, preclude the use of these models until a more in depth understanding of the pathogenesis of IVDD is obtained. Additionally, the use of a species highly regarded as a companion animal raises ethical concerns, which in the modern day cannot be overlooked.

4.2.3. Size and biomechanics

Scaling of specific parameters is relevant to model choice. The majority of models found in this review used animals with discs that were smaller in size compared to human discs. This is relevant in research involving mechanical testing considering the extensive use of mice and rats as preclinical animal models. A previous study that investigated the mechanical properties of mouse and rat discs identified that these relatively small IVDs possess similar mechanical attributes of compression and torsion alike to humans, validating them as a suitable mechanical model of the human disc. ¹¹⁶ Furthermore, a study comparing the normalized geometric parameters of disc height, width and NP area in numerous species found that the mouse and rat lumbar, and mouse tail discs are the closest geometrical representation of the human lumbar IVD. ⁹⁹

All animals, including humans, have greater axial rotation in the cervical spine compared to the lumbar region; however, lumbar range of motion is greater in humans compared to sheep, pigs and calves, particularly in flexion/extension. ⁸ In humans, the lumbar spine supports the upper body, while in quadrupeds it does not; however, the forces required to stabilize the spine horizontally may subject the quadruped spine to greater loads and therefore relate to the increased bone mineral density in some animals compared to humans. ⁸ Elliot and Sarver (2004) however, identified conclusive correlations between animal body weight and their lumbar spinal properties. ¹¹⁶ This finding supports not only the use of the most common animal species determined by this review, being rodents, but also the use of quadruped animal spines to mechanically model the human spine.

Alternative parameters that have been tested comparatively across species include the mechanical factors of torsion and axial compression, and the biochemical factors of collagen, glycosaminoglycan, and water content. ¹⁰⁰ , ¹⁰² Disc axial mechanics and the biochemical content of glycosaminoglycan and water were similar between species following normalization. ¹⁰⁰ This likeness in comparison was also evident for the parameters of torsion mechanics, except for sheep and pig discs which were statistically different from human discs, and collagen content, except for calf and goat discs for the same reason. ¹⁰² It should also be noted that in caudal rat and bovine tail discs, often utilized for mechanical degeneration models due to ease of access to the disc, the load transmission across the IVD joint likely differs considerably from that in humans, as they are non‐weight‐bearing structures and do not have any posterior elements or facet joints. ⁴⁰ , ¹¹⁷

4.2.4. Ethics

Ethical considerations and guidelines for animal use with respect to replacing, reducing and refining the use of animals in scientific research continues to improve. ¹¹⁸ Justification of species, method of acquiring degenerative disc disease and sample size is imperative prior to commencing an animal model study. This poses limitations on potential studies because as mentioned previously, despite some species displaying more similarities with human IVDs than others, they may have further restrictions for their use in research. Additional limits may be placed on the length of time that animals with induced IVDD can be maintained alive as their welfare is compromised due to pain and discomfort depending on the treatment. IVDD is a painful physiological process and for all preclinical animal models studying this disease process, animal welfare must be prioritized with euthanasia endpoints clearly outlined prior to commencing the study.

4.3. Methods of IVDD induction

All studies included in this review observed degeneration of the target IVD through either one of the many IVDD induction methods outlined in the results and appendix A, or by allowing for spontaneous disc degeneration to occur naturally. This comparative summary will serve as an appropriate consolidated guide for future researchers approaching studies in these respective areas of IVDD research and in selecting suitable species to model IVDD.

For surgical methods of IVDD induction, there was significant variation in the models reported. Needle‐puncture models as described by Lipson and Muir (1981) in the rabbit, ¹¹⁹ have evolved to include scalpel blades (annulotomy), ⁴⁶ , ¹²⁰ , ¹²¹ , ¹²² , ¹²³ , ¹²⁴ , ¹²⁵ , ¹²⁶ , ¹²⁷ , ¹²⁸ , ¹²⁹ , ¹³⁰ , ¹³¹ , ¹³² annulectomies, ¹¹² total discectomies, partial and complete nuclectomies, ⁸ nucleus pulposus aspiration ¹⁴ , ²⁰ , ¹³³ , ¹³⁴ , ¹³⁵ , ¹³⁶ and drill bit injuries. ⁷ , ¹⁶ , ¹³⁷ , ¹³⁸ Clearly, these surgical insults vary in the burden suffered to the animal and therefore, the degree of IVDD that is induced. Consequently, there is also variation in the effect that can be measured or assessed on the experimental endpoints of these studies. The researcher must have a sound scientific understanding of the result of the surgical insult. Also, the question of surgical experience and surgical skill needs to be considered, as these surgeries can be challenging, difficult to reproduce identically and require minute surgical approaches.

In rodents undergoing surgical intervention, degenerative characteristics were first identified as soon as 1 week postprocedure, with the most common timepoint of identifying disc degeneration being 2 weeks postsurgery (Table 2). Similarly in rabbits, the most common timepoint of identifying degenerative characteristics was 4 weeks postsurgery. These results highlight the efficiency of surgical models in inducing degenerative changes in IVDs, independent of animal species selection, and the reliability of such procedures as they can be methodically reproduced by researchers with adequate surgical experience. While surgery was the most common mechanism for inducing IVDD in the reviewed preclinical animal models, alternative methods included chemical induction and genetic modifications. There was a variety of chemical methods recorded by the reviewed papers for inducing IVDD (Figure 4B; Appendix A). The researcher should be well informed of the mechanism by which IVDD is induced with these chemicals, the possible side effects and if these criteria align with the objective of the study.

Designing an animal model to investigate disc degeneration must incorporate a multifaceted approach of critical decision making where all variables are accounted for and taken into consideration concurrently to produce the most appropriate model for the specific research question being asked. Models incorporating species that undergo spontaneous or natural IVDD due to aging are clinically most alike to the naturally occurring pathogenesis of IVDD as seen in humans. Spontaneous degeneration models, however, entail uncertainty about the origins and pathology of occurrence, and more specifically the stage of the degeneration pathway that each animal is at. Using an inductive model allows for a more controlled population from which treatment strategies can be more appropriately compared and thus experimentally induced degeneration models are considered by the authors to be ideal for creating accurate and reproducible animal models of IVDD. Needle puncture models are reliable and have been successfully used by numerous researchers to model disc degeneration, most commonly in rabbits and rodents. Any form of direct insult to an animal's disc that can be performed in a controlled manner is a reliable method of achieving degenerative characteristics like those seen with IVDD in humans. Similarly, in vivo surgical instrument implantation that alters the normal loading mechanics of a specific vertebral column section has proven effective to accelerate degeneration of target discs. Chemical and genetic methods too have proven useful in creating degenerative disc models. These models, however, require further investigation to discern if the gene alteration or chemical agent employed will have any significant effect on the therapeutic treatment option being studied. While surgical models require trained surgeons for procedure accuracy, they provide a reliable model for investigation into therapeutic treatment options of lower back pain in humans.

Furthermore, the studies reporting genetic modifications to induce IVDD and spontaneous IVDD, otherwise labeled as naturally occurring or aging, exhibited an equally similar extent of variation in model type. From the genetically induced IVDD studies, twenty different genes were targeted for testing of their role in IVDD across the sub‐classifications of gene knock‐in, knock‐out or over‐expression (see Results section and Appendix A). Only mice and rabbits have been used for genetic IVDD preclinical studies and degenerative characteristics were observed most at 8, 20 and 24 weeks of age in mice, and approximately two years of age in rabbits (Table 3). Genetic models may be a beneficial alternative to surgical interventions in establishing a minimally invasive reproducible IVDD model. An effective genetic knockout model of apolipoprotein E in rabbits has been produced; however, this is an indirect method of causing IVDD as it relies on elevating cholesterol and triglyceride levels in the cardiovascular system before any affects are seen in the IVDs. ⁵¹ From the spontaneously occurring IVDD studies there were a total of eight different species subject to analysis (see Results section and Appendix A). The drawbacks of naturally occurring animal models is that their availability depends solely on the rate of occurrence and their use for investigating response to treatment methods of IVDD is limited because the reason for disc degeneration remains unknown. ⁵ Artificial animal models have the added benefits of being experimentally reliable, reproducible and the biological degeneration is controlled, therefore enabling effective comparative analysis of these degenerative methods between studies. Experimentally induced disc degeneration has been extensively described in the reviewed literature and as such, the authors recommend prioritizing these methods for investigating human IVD responses to therapeutic intervention.

The studies analyzed in this review were generally undertaken to establish a reliable animal model that replicates IVDD in humans as best possible, while also exploring therapeutic interventions to halt, limit, or reverse the degeneration cascade or associated symptoms of disc degeneration in humans. Potential therapeutic treatments for IVDD that were explored by the papers incorporated in this review include the use of calcitonin, ¹³⁹ polyester amide microspheres, ⁹⁰ alendronate, ¹⁴⁰ syndecan‐4, ¹⁴¹ sulforaphane, ⁷² antioxidants, ⁷⁰ the transplantation of mesenchymal stem cells, ¹⁴ the implantation of a poly lactic‐co‐glycolic acid plug, ¹⁴² injectable polymethyl‐methacrylate and bovine collagen, ¹⁴³ and injectable discogenic cell therapy. ¹³⁵ No one animal model can truly replicate the degeneration and regeneration of the human IVD and therefore researchers must compromise to choose an optimal model that allows for the most suitable investigation of disc degeneration and treatment method that aligns with their research question.

4.4. Experimental endpoints of intervertebral disc degeneration

There was significant variation in the diagnostic endpoints applied by differing studies in determining the efficacy of their preclinical animal model for IVDD. Endpoints utilized vary from imaging modalities such as radiography, CT and MRI to histopathology, biochemical assays, immunohistochemistry, mechanical testing, and a variety of other measurable outcomes. ⁹ , ¹⁰ The variability observed in the broad categories of experimental endpoints was further broadened by the concurrent variation within these categories, such as the many different histological stains and biochemical assays used for histology and biochemistry respectively. A combination of experimental endpoints is often employed by researchers to investigate a multitude of outcomes and to best analyze the effectiveness of the IVDD preclinical animal model with some specific tests related to specific research questions. However, with a lack of consistency recognized amongst the reviewed studies, it gives rise to speculation as to what are the most suitable outcomes to be measuring and how can these be standardized for effective comparison of similar or differing preclinical animal models of IVDD.

Additionally, suitable, and reliable, endpoints are necessary to verify the onset of IVDD and to evaluate its progression appropriately. Our systemic review highlighted the great variety in the experimental endpoints employed in preclinical models. The endpoint choice is influenced by a multitude of factors. The researcher(s) must have a sound understanding of the preclinical animal model chosen, and the suitability of experimental endpoints to effectively evaluate the method of IVDD induction and especially, the effectiveness of any treatment strategies being evaluated.

4.4.1. Diagnostic imaging

MRI is the gold standard imaging modality for evaluation of the spinal cord, intervertebral disc, paraspinal soft tissue structures and other neurological structures. ¹⁰ On T2‐weighted MRI images of the spine, the healthy intervertebral disc exhibits high signal intensity in the inner water‐ and proteoglycan‐rich nucleus pulposus, with low signal intensity in the fibrocartilaginous annulus fibrous. ¹⁴⁴ For IVDD, there is typically a reduction in the T2 hyperintensity in the nucleus pulposus and changes in the tissues surrounding each disc (end plate sclerosis, vertebral osteophytes, and disk herniation). ¹⁴⁵ Analysis of MRI sequences can provide both qualitative (IVD morphology, endplate changes, etc.) and quantitative (nucleus pulposus size, hydration, proteoglycan content, etc.) data to further define IVDD.

Pfirrmann et al. developed an MRI grading system for the semiquantitative assessment of the human lumbar IVDD condition that is now the most widely used system. The grading system focuses on characteristic changes in the structure of a disc (T2‐weighted signal intensity, disc structure, ability to discriminate between the nucleus and annulus, and disc height). ¹⁴⁶ Increasing accessibility to MRI by researchers means there is increased use of this modality in preclinical animal IVDD models. High quality micro‐MRI images of rat spines using the Bruker Biospec 9.4 T MRI Scanner have been reported by Glaeser et al. ¹⁴⁷ ; however, most studies have utilized lower quality imaging with much lower field strength (1.0, 1.5 and 3.0 T). ⁹⁶ , ¹⁴⁸ , ¹⁴⁹ This limits the data that can be extracted from these studies for cross‐study comparisons, the detection of degenerative changes induced, such as identification of Modic changes in the vertebral bodies, and the reproducibility of MRI‐grading schemes such as the Pfirrmann scale and endplate change scoring.

Quantitative functional MRI techniques are a useful modality for the early assessment of IVDD in comparison to traditional MRI techniques. Multiparametric quantitative MRIs, unlike traditional MRIs that are used to measure gross tissue structure, facilitate the biological investigation of microstructural tissue composition and biochemical processes associated with the pathophysiology of disc degeneration. ¹⁵⁰ Sequences that have been used to investigate biochemical changes related to IVDD include T1‐ and T2‐weighted mapping, and diffusion weighted imaging (DWI). In recent IVDD research, these sequences have reliably indicated biochemical changes in the hydration status and proteoglycan content of degenerative discs in goats, ¹⁵¹ sheep, ¹⁵² pigs, ¹⁵³ and dogs. ¹⁵⁴ The use of these quantitative methods is gaining popularity in IVDD research due to the capability of the results in providing quantifiable comparison between study cohorts and for use in longitudinal studies. Furthermore, DWI proved to be an effective indicator of cervical disc degeneration in humans. ¹⁵⁵ With the prevalence of quantitative MRI methods increasing, the use of this modality in preclinical animal research presents an important consideration for prospective research to aid in the comprehensive characterization of degeneration.

Radiography was the second most used imaging modality in the reviewed studies (103/259, 39.77%). This is likely due to its reduced costs, increased accessibility, and the fact it requires the least amount of technical experience to both perform and review. Despite being unable to visualize IVDs on radiographs, a disc height index (DHI) percentage has been reported as an effective representation of IVD size for objective analysis, with the DHI being measured prior to and at certain time points after induction of IVDD in preclinical animal models. Radiographs can also highlight secondary degenerative osseous changes; however, with advancements in technology, other modalities such as micro‐CT offer a higher image quality for diagnostic capabilities and detecting more minute changes in the IVDs themselves and surrounding supporting structures such as the vertebral body and facet joints.

Each imaging modality has their advantages and limitations. Radiography is non‐invasive, efficient and can easily be conducted at several time points throughout the course of the IVDD animal model to assess progression. Whereas CT and MRI, while significantly more expensive, provide much higher quality images and are therefore more sensitive to detecting degenerative changes associated with IVDD. Micro‐CT allows for even higher resolution scanning of bone structure, which is required for analysis of small rodent spinal units. ¹⁵⁶ , ¹⁵⁷ Micro‐CT may reveal the osseous anatomy in more detail, including assessment of the internal osseous lamellar structure and vertebral endplates, thus offering a greater insight into osseous and mineralized degenerative changes associated with IVDD.

4.4.2. Histology

Histological features of the disc can be seen via different staining methods, with degenerative characteristics able to be graded by outlined classification systems. The histological grading scale established by Masuda et al. ¹²³ was commonly employed by the reviewed literature listed in Appendix A. This particular grading scale takes into account the following characteristics: extent of disruption to AF fibers, interruption to the border between the AF and NP, decrease in NP cellularity and condensation of the NP matrix. ¹²³ While commonly used, this was not the sole classification system used with the reviewed articles either incorporating this system, other systems or no system in particular. Culminating from a JOR Spine and ORS Spine Section collaboration to improve the consistency of histopathological grading, publications associated to this project recently proposed comprehensive IVD histological grading schemes for the mouse, ¹⁵⁸ rat, ¹⁵⁹ rabbit, ¹⁶⁰ and large animal models. ¹⁰ The grading schemes were developed following a literature review, survey from opinions of expert researchers and clinicians in spinal research and a validation study of experts and inexperienced IVD histology graders. ¹⁶¹ The authors intend for the proposed grading system to reduce variability and provide more objective comparisons of histological analyses between studies and preclinical animal models.

Hematoxylin and eosin (H&E) was the most common histological stain used, often accompanied by either Alcian‐blue or Safranin‐O/Fast green, for histological analysis. H&E and Alcian blue/Safranin‐O are useful baseline stains to identify cellular components (cytoplasmic/nuclear) of the disc and glycosaminoglycans respectively; however, due to the complexity of the disc compartments, they have limitations in delineating the IVD structures compared with their use for cartilage. ¹⁶² More recently, triple dye methods have been developed to increase histological resolution, such as Masson's trichrome. ¹⁷ , ¹⁶³ Picrosirius red has also been introduced for determining collagen fiber orientation, in ABPR (alcian blue‐picrosirius red) combination dye methods with hematoxylin or Safranin‐O for distinctive staining of proteoglycans (blue), collagen (red) and nuclear contents. ²⁹ , ⁸⁸ , ¹⁶⁴ , ¹⁶⁵ , ¹⁶⁶

4.4.3. Gross morphology

Gross examination of IVD's is another feasible method of grading degeneration. Through direct visualization or high‐resolution photographs, assessment of the principal structures of the IVD, being the NP, AF, and CEPs, can be made. The most employed gross morphological grading scale cited in the reviewed literature was published by Thompson et al. ¹⁶⁷ This type of macroscopic assessment has the benefits of being easy to conduct, is of low cost and is clinically relevant, as it was originally designed for assessment of human IVDs. Importantly, visual examination is inherently subjective, thus the variability in assessment should be minimized by using two independent assessors. ¹⁰ Additionally, as an experimental endpoint, it cannot be applied longitudinally in vivo unlike the imaging modalities of radiography, CT and MRI. While gross morphology is not appropriate for detecting small changes, it is a simple and easy technique that investigators can employ to gather additional comparative information from their research project.

4.4.4. Biochemical, gene and protein analysis

Studies in humans and dogs have determined IVDD to be associated with “inflammation, altered matrix synthesis, catabolic metabolism, cell death, and neural and vascular ingrowth in the disc and surrounding tissues”. ⁹ Biochemical, gene and protein analysis may assist in identifying certain metabolite products or alterations in protein expression, characterizing the process of IVDD and provide avenues for exploration into preventative or therapeutic targets. Quantitative biochemical assays commonly utilized in the reviewed literature (Appendix A) included the dimethylmethylene blue (DMMB) and hydroxyproline assays for measuring glycosaminoglycan content and collagen content respectively. While effective at highlighting content levels, these assays do not provide information regarding the quality of the matrix content so must be viewed prudently. To the knowledge of the authors, no reference levels exist for specific gene expression in the IVD of certain animal species. The challenges of gene expression profiling in the cartilaginous IVD includes low cellularity, difficult homogenization, and an ECM rich in glycosaminoglycans that can inhibit the real time polymerase chain reaction (PCR). ¹⁰ Gene expression levels can be unique to protein levels, thereby employing experimental techniques such as Western Blot analysis, enzyme‐linked immunosorbent assays, PCR and immunohistochemistry can reveal a much more profound analysis of IVDD at a cellular level.

Many different experimental endpoints can be employed when investigating IVDD. Each technique assesses a unique aspect of disc degeneration whether it be macroscopically, microscopically with histology or alternative staining methods such as immunofluorescence, biochemically or through advanced imaging modalities as discussed previously. To a lesser extent, other endpoints that can be employed include techniques such as biomechanical testing ¹⁰ and cell culture, ²⁴ , ³⁹ , ⁹⁶ , ¹³¹ , ¹⁶⁸ , ¹⁶⁹ as they offer diverse quantitative insight into varying aspects of disc degeneration. Investigating experimental protocols and their suitability to accurately assess IVDD is a focus for future research to assist with standardizing experiments to achieve relatability of results and increase translational applicability to human IVD health and disease.

4.5. Limitations of available models

Research into the human IVD is difficult as an in vivo model due to ethical considerations and government regulatory restriction. ⁸ Testing in preclinical animal models plays an important role in investigating the disc as they represent a critical surrogate in clinical translation with the purpose of furthering our understanding about the pathophysiology, treatment and prevention of IVDD. The ideal animal model should represent similar morphological and biomechanical properties to that of the human IVD ⁷ ; however, there are many fundamental differences between animal species and humans including but not limited to cell population, tissue composition, disc and spinal anatomy, biological development and functional biomechanical forces that the disc is subjected to daily. ⁸ While our understanding of the disc and it is degeneration has been enhanced by the exponential growth of spinal research in the past two decades, there are unavoidable disparities between animal species and humans and the results achieved from preclinical animal models need to be viewed with this knowledge in mind.

The advantages of experimental animal models, despite the aforementioned differences, comprise their availability and ability to be produced in large numbers, making them suitable candidates for developing reproducible disc degeneration as a surrogate model of human spinal research. ¹⁷⁰ Inducible disc degeneration involving a scalpel annulotomy and needle stab incision in a canine model was first published in 1948. ¹⁷¹ A diverse array of IVDD methods have since been explored in varying animals with researchers employing models that feature surgical or chemical disc injury, non‐invasive mechanical stress, genetic modifications and natural occurrence. ⁵ Similar variation can be observed in the type of animal model used and the experimental endpoints used to measure between control and experimental groups of animals in the reviewed papers. The lack of consistency in selected species, methods of achieving degeneration, and both experimental timepoints and measures that have been used in the available preclinical studies restricts the translatability between models and the extent to which comparisons can be made regarding the effectiveness of a model to successfully produce IVDD or of a treatment method in restoring IVD structure and function.

Finally, a drawback of focusing only on the macroscopic and microscopic parameters of disc degeneration restricts the clinical relevance of IVDD associated with lower back pain. Subjective and quantifiable behavioral and pain outcome measures, examples of which include gait analysis, species‐specific grimace scales, motion analysis and telemetry, may be more appropriate as a tool for assessing the relationship between degeneration and pain. ¹⁰ The majority of well validated measures of pain exist primarily for rodents, and not for larger animal models, so it is important to note that a compromise often exists between measuring pain or having a disc size‐scale that more closely represents humans. Pain behaviors and neurovascular ingrowth are important factors for consideration when assessing IVDD and characterizing these in each species used for in vivo models is an area requiring further research. While achieving clinical relevance, ethical concerns for animal wellbeing need to be appropriately addressed with experienced veterinarians to establish suitable endpoint criteria.

5. CONCLUSION

The findings of this review demonstrate continued improvement in the body of knowledge in relevant literature over time, supporting the interest in and utility of preclinical animal research into understanding the progression of IVDD and the effectiveness of therapeutic interventions. Although a wide range of species were treated with quite different interventions and assessed using disparate techniques of variable complexity and objectivity, the hallmarks of IVDD were discerned in all models. This speaks to the sensitive nature of the IVD and susceptibility to harmful stimuli as well as the detection ability of several different assessment techniques. Currently, no experimental animal model parallels human disc degeneration however many techniques have been trialed to produce a reliable model that can be used to study the degeneration and regeneration of the intervertebral disc.

The type of animal species coupled with their age are important considerations when designing a model of IVDD. It is fundamental that skeletal maturity has been achieved prior to commencing any induction method of degeneration in these animals, as the severity of IVDD in humans has been shown to correlate with age. Furthermore, it may be beneficial for investigators to address what rate and extent of degeneration is desired from the model, prior to selecting what they consider to be the most appropriate animal species. Finally, while acute treatments may be convenient from a time saving point of view, slowly progressing is likely to better match the human scenario but this needs to be balanced with economical and logistical considerations such as animal husbandry.

Researchers are well equipped to identify changes in treated discs when compared to healthy controls and therefore a comparison of the effectiveness of treatments including animal model and measurement modality would be a useful guide for future research. However, due to the different types of evaluation including qualitative and quantitative measures, grading scales that differ in scale and criteria, different time points and ages of animals used, this quickly becomes a futile concept. Recently published standardized scoring systems as part of the ORS Spine section initiative should be used for grading histopathology and diagnostic images of degenerative samples in future IVDD research. These collaborative and thorough guidelines will assist the spine research community to achieve consistency and translatability amongst prospective data analysis.

Additionally, incorporating multiple experimental outcomes that provide unique assessments of the IVD and the degeneration cascade only enhances the accuracy of data obtained from experimental work. CT for example is best used to assess bony endplate changes with less sensitivity for soft tissues, whereas MRI, depending on sequence used, is less useful for bone structure and optimal for assessing hydration relevant to proteoglycan content and inflammation. Despite assessing different aspects of IVDD, each technique contributes to creating a more complete understanding of the IVDD condition and implications of these changes. There is no one technique capable of detecting all changes, so investigators conducting research with animal models must have a clear idea as to what aspect of IVDD, or characteristics of degeneration, they are endeavoring to measure in their research.

We hope this review serves as a reference for researchers when considering species, method of inducing IVDD and experimental endpoint selection, and in doing so, will facilitate more reliable comparison of results between published studies. Further research is essential to continually improve our understanding of the pathophysiology of IVDD in the spine field. This review can assist researchers by providing a summary of a wide scope of in vivo preclinical animal models, acting as a guide for cross‐study comparisons and ultimately enhancing the translatability and clinical relevance of prospective research.

AUTHOR CONTRIBUTIONS

All authors provided substantial contributions to research design, acquisition, analysis, and interpretation of current literature. Onur Tanglay: search strategy, reviewed literature. Daniel Poletto and Matthew Pelletier: conceived the appendix table and descriptive statistics of the reviewed literature. Daniel Poletto and James Crowley: drafted the manuscript and provided supportive material for the various sections. William Walsh: study design and concept development. All authors have read and revised the whole manuscript critically and approved the final submitted version.

CONFLICT OF INTEREST

None to report.

Supporting information

Appendix S1: Supplementary information

Poletto, D. L. , Crowley, J. D. , Tanglay, O. , Walsh, W. R. , & Pelletier, M. H. (2023). Preclinical in vivo animal models of intervertebral disc degeneration. Part 1: A systematic review. JOR Spine, 6(1), e1234. 10.1002/jsp2.1234