The Relationship between Diabetes Mellitus Type II and Intervertebral Disc Degeneration in Diabetic Rodent Models: A Systematic and Comprehensive Review

The number of diabetic patients grows constantly worldwide. Many patients suffer simultaneously from diabetes mellitus type 2 (T2DM) and intervertebral disc disease (IVDD), suggesting a strong link between T2DM and IVDD. T2DM rodent models provide versatile tools to study this interrelation. We hypothesized that the previously achieved studies in rodents approved it. Performing a search in the publicly available electronic databases according to our inclusion (e.g., experimental study with clearly outlined methods investigating IVDD in diabetic rodent models) and exclusion (e.g., non-experimental) criteria, we included 23 studies from 1992 to 2020 analyzing different aspects of IVDD in diabetic rodents, such as on pathogenesis (e.g., effects of hyperglycemia on IVD cells, sirtuin (SIRT)1/p53 axis in the interrelation between T2DM and IVDD), risk factors (e.g., high content of advanced glycation end-products (AGEs) in modern diets), therapeutical approaches (e.g., insulin-like growth factor (IGF-I)), and prophylaxis. Regarding their quality, 12 studies were classified as high, six as moderate, and five as low. One strong, 18 moderate, and three mild evidences of the link between DM and IVDD in rodents were found, while only one study has not approved this link. We concluded that T2DM has a devastating effect on IVD, particularly in advanced cases, which needs to be further evaluated.


Basics
Low back pain (LBP) is one of the most common public health disorders in developed countries, with prevalence rates ranging from 12% to 35%. LBP affects up to 85% of people around the world and occurs at some point in their lives [1,2]. Degeneration of the intervertebral disc (IVDD) is considered to be an important underlying cause of LBP [3,4], back, neck, and radicular pains [5].
IVDs lay between the vertebrae in an alternating fashion giving to the spine its shape, flexibility, and high mobility in different axes and planes. IVDs represent up to 25% of the total spine length and act as distributors of loading and thereby resist compressive loads as well as tensile and shearing stresses [6,7]. The IVD is composed mainly of an external fibrous ring called annulus fibrosus (AF) and an internal jelly-like material called nucleus pulposus (NP), which in healthy discs appears like a whitish cotton-ball in healthy discs [8]. Both parts are connected by a transition zone. The NP has a hydrogel like consistency in healthy discs and becomes more fibrous in aged and degenerating IVDs [9]. AF attaches to the posterior longitudinal spinal ligament (PLSL) and fuses with the endplates of the upper and lower vertebrae. The AF consists of an external zone of concentric lamellar fibers of collagen type I and an internal zone of collagen fibers of type II. The localization of the NP within the IVD depends on loading conditions at the different segments (cervical, thoracic, or lumbar discs) [10,11]. Hence, differences among species exist as loading of the individual spine differs. In humans and rats, the NP in the lumbar spine is eccentrically and more dorsally localized between the central and posterior parts of IVD and exhibits properties of an incompressible water-filled cushion ( Figure 1). The IVD contains glycosaminoglycans (GAGs) that comprise mainly chondroitin-4-sulfate, chondroitin-6-sulfate, and keratan-sulfate in young and aged people and dermatan-sulfate in the elderly [4,6,8,[12][13][14]. In advanced aged humans, AF and NP begin to be dehydrated and partially collagenized [8]. However, in addition to aging, also metabolic diseases, such as diabetes mellitus, affect IVD homeostasis [5]. In this context, the studies of Fields et al. and Acevedo et al., showed that bone and IVD structure were indeed affected by type 2 diabetes mellitus (T2DM) in the University of California, Davis-Type 2 diabetes mellitus UCD T2DM rat model [15,16].
Cells 2020, 9, x FOR PEER REVIEW 2 of 32 and an internal jelly-like material called nucleus pulposus (NP), which in healthy discs appears like a whitish cotton-ball in healthy discs [8]. Both parts are connected by a transition zone. The NP has a hydrogel like consistency in healthy discs and becomes more fibrous in aged and degenerating IVDs [9]. AF attaches to the posterior longitudinal spinal ligament (PLSL) and fuses with the endplates of the upper and lower vertebrae. The AF consists of an external zone of concentric lamellar fibers of collagen type I and an internal zone of collagen fibers of type II. The localization of the NP within the IVD depends on loading conditions at the different segments (cervical, thoracic, or lumbar discs) [10,11]. Hence, differences among species exist as loading of the individual spine differs. In humans and rats, the NP in the lumbar spine is eccentrically and more dorsally localized between the central and posterior parts of IVD and exhibits properties of an incompressible water-filled cushion ( Figure  1). The IVD contains glycosaminoglycans (GAGs) that comprise mainly chondroitin-4-sulfate, chondroitin-6-sulfate, and keratan-sulfate in young and aged people and dermatan-sulfate in the elderly [4,6,8,[12][13][14]. In advanced aged humans, AF and NP begin to be dehydrated and partially collagenized [8]. However, in addition to aging, also metabolic diseases, such as diabetes mellitus, affect IVD homeostasis [5]. In this context, the studies of Fields et al. and Acevedo et al., showed that bone and IVD structure were indeed affected by type 2 diabetes mellitus (T2DM) in the University of California, Davis-Type 2 diabetes mellitus UCD T2DM rat model [15,16].

Diabetes Mellitus and Musculoskeletal Disturbances
DM represents a chronic degenerative disease, which results either from partial or complete insulin deficiency or cellular resistance against insulin receptors in target tissues [18,19]. DM is classified into four major types-T1DM, T2DM, gestational diabetes, and specific DM. T1DM and T2DM are the most common types and are discussed below, while specific or secondary diabetes and gestational diabetes are less common. Gestational diabetes mellitus (GDM) is defined as any glucose intolerance by the onset or during the course of pregnancy, regardless of whether it resolves or remains after delivery. Specific types of DM could be caused by genetic defects of β-cell function or insulin action or could be secondary to diseases of the exocrine pancreas, such as endocrinopathies, drug abuses, and/or to toxic effects of chemicals, infections, autoimmune disorders, or secondary to genetic syndromes associated with DM [20,21]. DM is classified into two main types: DM type I that is caused mainly by the reduction of insulin secretion from β-cells located in the Langerhans islets in the pancreas or from the destruction of those cells because of an autoimmune or non-autoimmune reaction, while DM type II results from a deficient or absent cellular response to insulin at the receptor level. DM type I represents about 10-15% of DM and occurs mainly in children, while DM type II represents 85-90% and is observed predominantly in adults and elderly people, especially in the case of obesity [19,22]. DM is an international major public health problem which is expected to affect 300 million people by 2025 worldwide [18,22,23] and more than 360 million by 2030 [22]. DM is characterized by wide disturbances in carbohydrate, lipid, and protein metabolic regulation, leading to an extensive and long-term dysfunction and various organ failures of the eyes, kidneys, liver, ears, cranial and peripheral nerves, heart and blood vessels, in addition to musculoskeletal tissues (bones and cartilages) [5,18,19,24]. In general, it mainly affects connective tissues [25].
DM is associated with many musculoskeletal disorders, such as Frozen shoulder, Dupuytren's contracture, carpal tunnel syndrome, stiff hand syndrome, flexor tenosynovitis, joint stiffness, Charcot joint, gouty arthritis, osteoarthritis, rheumatoid arthritis, diabetic amyotrophy, diabetic muscle infarction, diffuse idiopathic skeletal hyperostosis, osteoporosis and osteoporosis-related fractures [22,25,26], and IVDD [5]. The underlying pathogenesis of joint stiffness is thought to be the increase of collagen cross-linking due to the non-enzymatic glycosylation of collagen with advanced glycation end-products (AGEs) formation [27]. However, AGEs were found in the NP of IVDs, mediating a dysregulation of aggrecan synthesis and IVD stiffening. Hence, their contribution to IVDD has been strongly suggested [3].

Diabetes Mellitus and Intervertebral Disc Degeneration
As many as 40% of cases of LBP go along with IVDD and upcoming evidence suggests that T2DM substantially contributes to the severity of IVDD [5]. The general pathogenesis of IVDD occurs by the loss and degradation of the disc's extracellular matrix (ECM) components/molecules, such as proteoglycans, collagen, GAGs, and water content resulting from an imbalance of the catabolic and anabolic mechanisms, leading to wide histological and biochemical changes, such as increased levels of pro-inflammatory cytokines and associated enzymes [3,4,13]. Furthermore, the nutritional routes of the IVD and cellular viability and activities of the disc will be insulted [4,13]. Lifestyle factors, such as smoking and obesity, as well as background diseases, such as T2DM, contribute to IVDD. These effects are most likely mediated by malnutrition at the tissue level [28]. Krishnamoorthy et al. reported that the continuous accumulation of AGEs, which are associated with hyperglycemia in T2DM, was responsible for IVD stiffening and the following destructive cascade [3]. The T2DM-associated biochemical alterations lead to biomechanical and functional disturbances, followed by proceeding IVDD impairing in the whole spinal column as a functional unit. Hence, grading systems to assess T2DM-associated IVDD severity were developed. In 1990, Thompson et al. provided a macroscopical grading system [17]. More recently Raj [4] graded IVDD as follows: (Grade 0) normal unchanged nucleus; (Grade 1) annular tearing limited to the inner area of AF; (Grade 2) annular tears, destroying the entire disc architecture, but the outer contour of AF remains unaffected; (Grade 3) AF and PLSL are completely disrupted, including deformity of the entire disc [4]. A computer tomography (CT)-grading of IVDD was first stated in the 1990s, then initially modified by Aprill and Bogduk et al. in 1992 [29] based on MRI and finally modified in 1996 by Schellhas et al. [4,30,31]. Pfirrmann et al. developed in 2001 an MRI-based five-graded classification system of IVDD of the lumbar spinal discs, which comprises MRI signal intensity, disc structure, the distinction between AF and NP, and disc height on MRI scan [32]. Because of its high ambiguity, subjectivity, and non-applicability for young and old patients, the Pfirrman grading system was criticized by Xiong et al. [33], and by Griffith et al. [34], who provided/suggested a modification based on the widening of the grading of the IVDD from 5 to 8 categories and the increase of its power by including the elderly spines [33,34].

Rodents as Experimental Diabetic Models in IVDD
Since the pathogenesis of IVDD remains unclear, animal models are widely used as versatile tools to elucidate different aspects of IVD degeneration and to study its progression in detail. Jin et al. mentioned that the selection of the ideal animal model for the investigation of IVDD was more difficult than expected, and the following points should be considered: the complicated and partly unknown etiology and pathology of IVDD, the need of similarities of anatomical and pathological features of IVDD to those of humans, the reliability, the reproducibility, the labor efficiency, and, lastly, the costs invested in the experimental groups in regards to the size required to get statistically significant results [35].
A couple of detailed reviews regarding the used animal models in studying of IVDD degeneration and IVD regeneration (IVDR) exist already [35][36][37]. Such animal models used to study IVDD and IVDR were rodents, dogs, sheep, rabbits, goats, and primates [37]. Daly et al. mentioned numerous factors like the persistence of notochordal cells and the disc size, geometry, and mechanical forces that should be considered in the selection of the best-suited animal model to study IVDD and IVDR [36]. Rodents are considered to be a good choice model not only because of the low cost that allows statistically valid group sizes. Rodents could be a strong model, and the rat and mouse tail provide optimal access to IVDD, e.g., with the use of mechanical injury, asymmetrical compression, or administration of digestive enzymes. The advantages of rodent models are also the availability of numerous models of T2DM and obesity, the ease to manipulate them genetically, the short breeding span, the access to physiological and invasive testing, and the balanced cost-effectiveness [38,39]. In contrast, limitations include the fact that monogenic models are not representative of most human disorders [38]. In addition, there are other disadvantages, such as the persistence of notochordal cells in rodents, the obvious disc size discrepancy between humans and rodents ( Figure 1), biomechanical differences in mechanical loading, and the ethical obstacles in the case of, e.g., creating bipedal mice [36] to mimic human conditions. During prenatal life, vertebrae and IVDs of all mammals arise from aggregation of the mesenchyme around the notochord and the following segmentation. In humans, the notochordal cells start diminishing after birth and completely disappear at adulthood, while in other species, including rats and mice, the notochordal cells distinctly remain until adulthood [35]. The persistence of notochordal cells limits the adequacy and relevance of implementation of rodents as the animal model of choice in the trials, which investigate cellular regeneration therapies of IVD because the resulting IVDR could not be correctly evaluated, whether it arises due to the applied therapies/agents or due to the potentially preexisting notochordal cells [36]. Genetically modified mice models are used to investigate the role of specific proteins in the etiology of IVDD pathogenesis [36]. Many methods are reported to create IVDD in rat or mouse tail. Such are genetic predisposition (through mutation creation) [35,37], mechanical loading (application of altered mechanical stresses by bending, cyclic chronic compression, and postural changes) [36,37], chemical digestion (through chemonucleolysis by using chemical agents or enzymes, which cause pathophysiological IVDD, such as chymopapain and chondroitinase ABC) [36,37], and physical or structural disruption (through surgical interferences as puncture and annulotomies) methods [35,36].
Numerous methods [40,41] on how to induce T2DM in the experimental rodent models have been reported, and the majority of them are mice models. Fajardo et al. proposed a novel classification system of the used diabetic rodent models of T2DM based on the following criteria: "(A) spontaneous or diet-induced, (B) mono-or polygenic etiology, (C) obese or lean body type, and (D) by the timing of T2D onset, either before or after skeletal maturity" [40].
T2DM rat models are monogenic obese models (Zucker diabetic fatty (ZDF) rats), polygenic obese models (Otsuka Long-Evans Tokushima fatty (OLETF) rat; Israeli sand rat (ISR)), induced obesity by high-fat diet (Nile grass rat and desert gerbil), and non-obese models (Goto-Kakizaki (GK) rat; spontaneous diabetic Torii (SDT) rat) [42]. The ISR [43] is an early onset obesity model used before and after skeletal maturity. It develops hyperinsulinemia and frank diabetes. No reports on its musculoskeletal disturbances were found. The ZDF rat is a model of moderate obesity and has a leptin receptor deficiency. Only the male individuals develop early hyperinsulinemia and later hyperglycemia, which is followed by decreased insulin. In comparison with controls, it shows decreased femoral length and diameter, in addition to bone formation in lumbar vertebrae. The GK rat is a spontaneous diabetic rat model that maintains glucose levels below 200 mg/dL. It has a shorter femur and smaller lumbar vertebral height, in addition to lower femoral and lumbar bone mass density (BMD) [42]. The spontaneous diabetic Torii (SDT) rat is a non-obese model of T2DM characterized by a blood glucose level of about 600 mg/dL at the age of 24 weeks and low insulin secretion caused by β-cell dysfunction. By advancing age, it exhibits decreased femoral and tibial BMD, reduced tibial three-point bending stiffness and peak load, and reduced bone formation rate [38,40]. The Zucker diabetic Sprague Dawley (SD) rat is a genetic-modified rat model through a leptin receptor mutation and develops rapidly a frank T2DM. It exhibits reduced femoral length and diameter, diaphyseal cortical thickness, and cortical volumetric BMD, biomechanical changes, and reduced vertebral BMD [38,40]. The induction of T2DM in SD rats using alloxan has the following drawbacks: high mortality rates, ketosis, and T2DM is reversible, while the use of streptozotocin (STZ), which is used mainly for the induction of T1DM, can also induce T2DM, having the following advantages: higher selectivity to β-cells, lower mortality rate, and T2DM is irreversible [44].
The OLETF rat is a polygenic rat model of moderate obesity that could be used in diabetic and prediabetic experiments. The musculoskeletal changes of the OLETF rat model are poorly reported [38][39][40]45]. In this context, the studies of Fields et al. and Acevedo et al. showed that bone and IVD structure was indeed affected by T2DM in the UCD T2DM rat model [15,16].
The C57Bl/6J mouse is the optimal-studied model of diet-induced obesity; however, it could be an acceptable candidate to study DM. The mice do not develop frank T2DM, but rather a mild and transient type. However, it shows high-fat diet/obesity-related musculoskeletal changes, such as lowering of trabecular bone volumes and reduction of bone formation [46]. The MKR exhibits frank hyperglycemia, insulin resistance, and hyperlipidemia by the second week of age. It shows musculoskeletal changes, such as decreased femur stiffness and osteoblastic activity [40]. The Ob/Ob mouse is a model of severe obesity and develops obvious hyperglycemia, -insulinemia, and insulin resistance. It shows a low bone mass in the lumbar vertebrae and the long bones and disturbed physio-biomechanical properties, in addition to shorter femurs in comparison to non-diabetic ones [40]. The db/db mouse is a model of severe T2DM. Its musculoskeletal changes are not clearly addressed; however, disturbances in bone structure and reduction of the length of long bones are reported [47]. Diabetic manifestations are more obvious in males. Nevertheless, the severity of diabetes in both the ob/ob and db/db mice models is dependent on the genetic background: the black Kaliss's (BKS) mice background results in severe DM and early death. While on the C57BL/6 background, DM is mild and transient [48]. The yellow Kuo Kondo mouse develops severe obesity, insulin resistance, hyperglycemia, and -insulinemia by 8 weeks of age. Its musculoskeletal changes have not been investigated yet [40]. The TallyHo mouse is a model of early-onset, naturally occurring obesity and T2DM. Males of this diabetic model show reduced BMD of the trabecular bone at the distal femur [40]. The M16 mouse is a model of early-onset obesity and shows moderate hyperglycemia [40]. No investigations regarding its musculoskeletal changes were found. The NSY mouse is a model of mild obesity and shows increased glucose levels and hyperinsulinemia [40]. However, no reports of musculoskeletal changes are available. The TSOD develops obesity, impaired glucose tolerance, and hyperinsulinemia [40]. The occurrence of musculoskeletal changes has not been published yet. The NZO mouse exhibits impaired glucose tolerance and could develop frank diabetes. No studies regarding its musculoskeletal changes are described [38][39][40]45].
In contrast to mice, rat models allow more clinical handling and manipulation, so the researcher can repeatedly drain enough blood without affecting the animal, and more tissue samples are available for multiple downstream analysis.

Study's Basic Idea
We hypothesized that the previously achieved experimental studies in rodent models approved the positive link between T2DM and IVDD, particularly in rodent diabetic models. We undertook a systematic and comprehensive literature review, considering only the published experimental studies, which discussed the link between T2DM and IVDD in rodent diabetic models. Thereby, we also summarized and discussed the different rodent models used. We aimed to achieve the following objectives: (I) achieving a systematic comprehensive literature review of the experimental trials undertaken in diabetic rodent models to study the link between T2DM and IVDD, (II) revealing the strengths and weaknesses of the published studies, (III) combined and comprehensively displaying of the results of the achieved studies in the form of a critical constructive report, and (IV) identifying the unanswered questions in the currently available literature.

Data Search: Literature Sources
We conducted our systematic review and data search and selection according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [49]. We performed an extensive systematic electronic literature search in the available electronic medical databases, including PubMed, PMC, Cochrane Library, Wiley Online Library, Hindawi Publishing Corporation, Hindawi Journals, Web of Science, and Excerpta Medica dataBASE (EMBASE). We used separated keywords regarding the topic, results, methods, etc., such as DM, IVDD, IVD, rats, rat models, spine degeneration, cytokines, NP, AF, GAG, etc. and also titles, such as DM and IVDD, IVDD in rats, animal models in IVDD, etc. After the initial literature research, we performed a records screening, trying to find every relevant record or report from the obtained studies. The list of references for the collected papers was carefully checked to ensure the eligibility of the studies and their accordance with our topic. Each step of the above-mentioned three steps was separately performed by two reviewers (M.M.) and (M.K.), and after that, they were revised and checked by the third and fourth reviewers (A.A.) and (G.S.-T.).

Characteristics of the Initially Obtained Literature
We collected the published studies (9511 papers), which were identified as related to the topic (IVDD). Out of these, 2144 manuscripts were related to the topic IVDD in rodents, 1528 publications related to the topic IVDD in DM, and 27 publications related to our topic IVDD in diabetic rodent models, see Table 1 and Figure 2.

Literature Filtration and Selection
The studies that were included in this review were experimental and had three axes "IVDD, DM, and rodent models". Therefore, 9485 manuscripts, which were not relevant to our topic, were excluded.
They were either not applicable for this review, or they were not experimental. Finally, after accurate filtration, only 23 publications were included, which were relevant to our topic, see Tables 1 and 2 and Figure 2.

Literature Filtration and Selection
The studies that were included in this review were experimental and had three axes "IVDD, DM, and rodent models". Therefore, 9485 manuscripts, which were not relevant to our topic, were excluded.
They were either not applicable for this review, or they were not experimental. Finally, after accurate filtration, only 23 publications were included, which were relevant to our topic, see Tables 1 and 2 and Figure 2.

Study's Admission Requirements
In order to be included, the study should minimally fulfill 5 inclusion criteria, and no exclusion criterion ( Table 3). The inclusion criteria were the following: (I) the experimental study should have clearly outlined methods, which investigated effects of (II) T2DM on (III) IVDD in (IV) diabetic rodent models and focused on the pathogenesis, risk factors (RFs), treatment or prophylactic agent of the issue of interest, (V) the publication must be accessible in English with its record and appendices if present. The exclusion criteria were: (I) the study was not experimental (review, clinical, case report, etc.), (II) deficient of DM or (III) IVDD or (IV) rodent model, or (V) its methods were not applicable, unclear, or insufficient. Table 3. Inclusion and exclusion criteria.

Inclusion Criteria (V/V) Exclusion Criteria (I-V/V)
Experimental with clearly described methods (I) Non-experimental (review, clinical, case report, etc.) (I) T2DM (II) Deficient of DM (II) or IVDD (III) or rodent model (IV) IVDD (III) Unclear methods (V) Rodent models (IV) Study of pathogenesis, RF, treatment, prophylactic agent Available in English (V) Remark: to be included, the selected papers should fulfill five criteria (V/V) Remark: one criterion up to four criteria is/are enough to exclude the papers (I-V/V) DM: diabetes mellitus, IVDD: intervertebral disc degeneration, RF: risk factor.

Data Extraction
The selected studies (23 manuscripts) were also inspected regarding date of publication, publication origin, the focus of study (pathogenesis, risk factor, therapy, and prophylaxis), author details, the aim of the study, methods, type of used rodent model, results, etc. (see Table 4). The included studies were accurately and in detail examined by two reviewers (M.M.) and (M.K.), and the collected data were revised by two leading reviewers (G.S.-T.) and (A.A.).

Evaluation of the Studies
We aimed to evaluate each study regarding its quality and strength by the individual and combined assessment of its components, such as aim, methods, results, etc. In order to achieve this, we adhered to the Newcastle-Ottawa scale (NOS) [50,51]. However, the NOS is highly efficient as an evaluative and analytical tool in the systematic reviews involved in epidemiological and clinical studies, such as cross-sectional studies, cohort studies, etc. Yet, we had to generate our own system, which is a point-based assessment system adapted to experimental studies and applied it in this review (see Section 2.7 and Table 5).

Study Scoring System (SSS)
This system evaluated each study according to 5 components (abstract, aim, research question (RQ), methods, and results). For each one of these components, there were assessment tools/parameters, see Table 5. There were five (5) criteria for abstract evaluation, which comprised clarity, conciseness, readability, completeness, and outline; five (5) criteria for RQ, which were feasibility, interestingness, novelty, ethics, and extent of relevance to our issue; five (5) criteria for aim evaluation were realism, experimental feasibility, clinical applicability, and profitability; five (5) criteria for methods evaluation were appropriateness, efficiency, experimental availability, comprehensiveness, and correct achievement, and five (5) criteria for results evaluation were reliability, absence of bias, validity, applicability, and measurability. The SSS comprises 25 points, 5 points for each item/component. The quality assessment (QA) of each study depends on the gained points in the SSS, as follows: high quality (21-25 points), moderate quality (16-20 points), and low quality (≤15 points).

Evidence Assessment System (EAS)
In addition to the above-described assessment system, we set an additional evaluative parameter to estimate the strength of the detected evidence of DM influence on IVDD (rating in the range from 1 to 8), see Table 5. This system depends on the total sum of the different types of findings. Each finding for an interrelation between DM and IVDD was an indicator and the summation of all indicators, supporting that the hypothesis "DM and IVDD are interrelated in the rodent model" exhibits the final/real evidence of the study. The supporting findings or indicators could be acquired by histopathology (H), immunodetection (I), molecular biology (M), biochemical (BC) or biomechanical methods (B), imaging (IM), clinical (C), and/or statistical (S) approaches; coincidence between the results, underline the relation between IVDD and T2DM in the used rat model-derived tissues and if included, also in human tissues. The evidence strength of IVDD and DM interrelation was documented as following: strong (≥5 indicators), moderate (3-4 indicators), mild (1-2 indicators), and absent (0 indicators). There were nine studies included, which were exclusively undertaken in vitro, and two showed, in addition to in vivo, also in vitro results and described the devastating effect of glucose in culture medium on IVD cells, see Table 4 and Table 5. In these studies, the IVD cells were isolated from non-diabetic rat models, and afterward, they were cultured in growth mediums with elevated glucose content of different concentrations. The rest of the included studies demonstrated the effect of DM on IVD tissues in vivo, as the IVD cells were isolated from diabetic experimental rat models and non-diabetic rat models as a control, see Table 4. Both in vivo and in vitro proceedings, were considered in our study to provide evidence for the link between DM and IVDD with different strengths of the detected evidence of DM influence on IVDD degrees according to the scoring system described above.

Identified Studies
We identified 9511 published studies related to the issue IVDD. Out of these 27 manuscripts were related to the topic IVDD in diabetic rodent models, which is our research field of interest [3,15,43,47,[52][53][54][55][56][57][58][59][60][61][62][63][64][65][66][67][68][69][70], see Tables 4 and 5. By further evaluation another four publications were excluded; two of them discussed the diabetic animal models used in IVDD, including rodents, but they were not related to IVDD. They focused only on studying the different animals, such as goats, rodents, dogs, etc., which were used as experimental diabetic models [35,36]. The third one was excluded because it discussed the role of adiponectin, mainly found in obese individuals, but it did not address any DM model [71], and the fourth excluded study dealt with a spinal injury in diabetic rats as one unit and not with IVDD as the main subject of this study [72]. After the final careful filtration, 23 studies were selected and included in our review, which were fulfilling the inclusion criteria listed in Table 3.

Origin of Included Studies
Regarding  Table 4 and Supplemental Figure S1A. According to their origin, eight studies were derived from China, representing 34.8% of total included studies, seven studies from Korea, representing 30.4%, five studies from the USA, representing 21.7%, and 3 studies from the USA with Israel, Taiwan, and Japan, representing 4.34% for each country, and no study from Europe, see Table 4 and Supplemental Figure S1B.

Data Assortment, Analysis, and Evaluation
We studied the collected publications to get the information regarding their focus of the trial, type of evidence, quality of the study, and evidence strength, see Table 5 and Supplemental Figure S1. We found that 14 studies representing 60.9% of the whole included studies in the review performed in vivo experiments in order to test the effect of DM as a metabolic disease on the IVD tissues and whether it is playing a role in developing IVDD using rodent diabetic models. Nine (9) studies, representing 39.1%, were undertaken in vitro and investigated the effect of different glucose concentrations on IVD tissues extracted from non-diabetic rat models. In addition, two studies combined in vivo and in vitro analyses. The controls in the in vivo performed studies were non-diabetic rodent models, while the controls in the in vitro performed experiments were IVD tissues cultured in normal or low glucose concentrations containing growth media, see Table 4. The in vivo studies were using mice (5×) or rats (9×), whereas all in vitro studies included were based on rat-derived IVD cells.
Fourteen studies, investigated the overall pathogenesis of DM in the development of IVDD, see Table 4, [53][54][55][56][57]59,61,[63][64][65]67,69], while two studies investigated a specific pathogenesis element [58,66], such as the role of Sirt1/p53 axis in diabetic IVDD and the effect of metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) on apoptosis promoted by high glucose in rat cartilage endplate (CEP) cells. The other four studies (17.4%) investigated the effect and pathogenesis of risk factors [3,52,60,70], such as the degenerative effect on spine/IVD tissues induced by chronic ingestion of AGEs and the long exposure to high glucose concentrations. Two studies explored preventive methods/agents (prophylaxis) of IVDD induced by DM, such as vitamin D and anti-AGEs drug [62,68]. One study (8.7%) surveyed the physio-biomechanical properties of the diabetic degenerated IVD [15,54]. One study investigated the pathogenesis of the diabetic IVDD by carrying out a therapeutic trial using IGF-1 injection as an anti-DM drug [47], see Table 4.

Quality of the Studies Included
According to the assessment of the quality of the included studies using our generated SSS, 12 studies, representing 52.2%, were graded as high in quality, six studies, representing (26.1%), were graded as moderate in quality, and the last five studies (21.7%) were graded as low in quality, see Table 5.
Most data supporting the link between IVDD and T2DM were statistically affirmed and achieved by immunodetection, followed by data based on biochemical and histopathological findings. Moreover, the in vivo studies exhibited one clinical and two biomechanical positive findings. The in vitro studies separately investigated the degenerative changes caused by T2DM in most of the anatomical structures of IVD (AF, NP, CEP, and notochordal cells).

Evidence Assortment and Assessment
Evidence was calculated based on the number of separate experimental approaches within one study, which revealed significant differences between the T2DM group and controls in regard to IVDD development. All studies based on diabetic rodents proofed the diabetic status of the animals by checking the blood glucose levels, mostly measured after fasting (Table 4). Regarding the type of found evidence (Table 5) (Tables 4 and 5). Nearly all in vivo studies presented histopathology. Since the biomechanical findings were limited to the included in vivo studies, only a few biomechanical investigations were performed [3,43]. If performed, imaging aimed to elucidate bone alterations accompanying IVDD in DM. Clinical features existed in only one included study (withdrawal test to assess pain) [69]. In vitro studies generally could not show histopathology, clinical evidence, and supporting imaging results. Therefore, the in vitro evidence was mostly found on methods based on immunodetection and biochemical assays. Nevertheless, there was a bulk of evidence deduced from the included in vitro studies. Molecular biological techniques were less often applied as a means to achieve evidence for IVDD caused by T2DM [47,59,69].
According to our generated Evidence Assessment System (EAS), see Table 5, evidence was defined in this review as the collective strength of the whole positive findings of each study by different methodological approaches. Four studies were evaluated, possessing strong evidence, representing 17.39% of the total included studies. Seventeen studies were evaluated to possess moderate evidence, representing 73.91% of the whole included studies. One study was evaluated to show low evidence, representing 4.34% of the whole included studies. Nonetheless, no studies showed only one positive finding or negative experimental finding, and only one study (4.34%) reported the absence of evidences [70]. -Expression of collagen type I in the DM group was higher than in controls in contrast to collagen type II -vWF was expressed in both, but was low in the DM group -MVD of the DM group was smaller compared to that of the controls -The apoptotic index (AI) in the DM group was significantly higher compared to that of the controls -Negative correlation between the MVD of EP and the AI of notochordal cells -Compared to controls, the EP MVD and the vessel width decreased or disappeared in DM rats  One diabetic and two non-diabetic groups (gender not mentioned): -Six-month-old lean SD rats ("control"), -obese SD rats ("obese"), -UCD-T2DM rats ("diabetic"); n = Two rat groups: 6-month old male OLETF (diabetic) and LETO (control) rats (10 per each group) -Rat DM model -Glucose tolerance tested -Tissue harvest: lumbar spine -Histopathology -TUNEL assay (apoptosis: AF cells) -Western blot: MMP-1, -2, -3, -13, tissue inhibitor of metalloproteinase (TIMP)-1, -2 -PCR: IL-1, -6 and TNF-α -OLETF rats showed increased body weight and abnormal 2-h glucose tolerance tests compared to LETO rats -The AI and the degree of Fas expression by AF and the -Expression of MMP-1, -2, -3, -13, TIMPs-1 and -2 was statistically higher in OLETF rats -Expression of IL-1, -6 and TNF-α was statistically higher in OLETF rats -Histological analysis showed more severe fibrosis and loss of lamellar pattern in AF tissues of OLETF rats  Three groups: -wild-type male C57BL/6J mice, -leptin receptor gene knockout, db/db mice -IGF-1 group Only male mice were included because of the DM protective effect of the female sex steroids -Mice DM model -Fasting blood glucose -IVD harvest (coccygeal + lumbar) -µCT -Histopathology -Immunohistochemistry: MMP-3 -Western blot: leptin receptor -Tunnel assay: apoptosis -PCR: sex-determining region SRY of the Y chromosome (SOX9), aggrecan, MMP-3 -Blood glucose levels were significantly higher in the db/db mice -T2DM in db/db group showed an association with significantly decreased vertebral bone mass and increased IVDD when compared to WT mice -db/db mice showed a higher percentage of MMP-3 expression and cell apoptosis than wild type mice -IGF-1 treatment partly reversed the findings -µCT: detection of IVD disc height index (DHI) -Biomechanical examination -Dietary effects on bone structure in Db/Db mice were sex-dependent and evident in females but not males -IVDs of female (but not male) Db/Db mice exhibited morphological changes, but no IVDD -Leptin receptor deficiency did not cause IVDD in 3 months old mice -DHI was not changed in any group -No biomechanical changes, except diminished torsional properties in leptin deficient mice.  Only one non-diabetic group (four-week-old male SD rats), from which the NPs have been harvested before exposed to hyperglycemic or normoglycemic conditions. Only one non-diabetic group (four-week-old male SD rats), from which the NPs have been harvested before exposed to hyper-or normoglycemic conditions.  One group of 24-week-old male SD rats, from which the NP cells have been harvested before exposed to hyper-or normoglycemic conditions.  A group of three 6-month-old male SD rats, from which CEPs were harvested -Negative control

Discussion
The literature research identified a plethora of published papers related to the topic of IVDD, from which 23 studies connecting to T2DM fulfilled the inclusion criteria. The majority of them (13) was performed in vivo, and the smaller section was based on in vitro testing. A fewer number of studies combined both results from in vivo and in vitro experiments [58,63]. The in vitro studies allowed separate investigation of rodent IVD-derived cell types, such as AF [55], NP [65], notochordal cells [54,59,60], or even EP chondrocytes [66,67], in regard to the link between IVDD and T2DM. Strictly speaking, notochordal cells represent precursors of the NP cells characterized by a specific expression profile and associated with multipotent differentiation and self-renewal capacity and, hence, regenerative potential [5]. The studies related to notochordal cells included here did not further address this specific expression profile, and hence, their discrimination from NP cells remains elusive [54,59,60]. Looking at the in vitro studies when high glucose was used to mimic hyperglycemia conditions, the concentrations applied substantially differ (e.g., 25 mM [65,66]) versus 100, 200, or even 400 mM [54][55][56]59,60] or 5.5, 25, 50, 100, and 150 µM [58]. In this respect, one has to face the fact that the blood glucose levels in diabetic rats (ZDF rats, 4-8 weeks) or mice were between 20 and 30 mM [47,73].
Most of the included studies focused on different aspects of the overall pathogenesis of IVDD in diabetic rodents (e.g., [52][53][54]). Besides the presence of elevated levels of advanced glycation end-products (AGEs) as a result of high glucose concentrations [3,62,63], they hypothesized a deficit in IVD nutrition caused by changes in the thickness and porosity of the bony vertebral EP and, subsequently, reduced microvascularity and vessel diameters in response to T2DM (Figure 3, [53]) as a trigger of IVDD development. Malnutrition also leads to enhanced expression of hypoxia-inducible genes in the IVD [15] (Figure 4). Cells 2020, 9, x FOR PEER REVIEW 23 of 32

Discussion
The literature research identified a plethora of published papers related to the topic of IVDD, from which 23 studies connecting to T2DM fulfilled the inclusion criteria. The majority of them (13) was performed in vivo, and the smaller section was based on in vitro testing. A fewer number of studies combined both results from in vivo and in vitro experiments [58,63]. The in vitro studies allowed separate investigation of rodent IVD-derived cell types, such as AF [55], NP [65], notochordal cells [54,59,60], or even EP chondrocytes [66,67], in regard to the link between IVDD and T2DM. Strictly speaking, notochordal cells represent precursors of the NP cells characterized by a specific expression profile and associated with multipotent differentiation and self-renewal capacity and, hence, regenerative potential [5]. The studies related to notochordal cells included here did not further address this specific expression profile, and hence, their discrimination from NP cells remains elusive [54,59,60]. Looking at the in vitro studies when high glucose was used to mimic hyperglycemia conditions, the concentrations applied substantially differ (e.g., 25 mM [65,66]) versus 100, 200, or even 400 mM [54][55][56]59,60] or 5.5, 25, 50, 100, and 150 µM [58]. In this respect, one has to face the fact that the blood glucose levels in diabetic rats (ZDF rats, 4-8 weeks) or mice were between 20 and 30 mM [47,73].
Most of the included studies focused on different aspects of the overall pathogenesis of IVDD in diabetic rodents (e.g., [52][53][54]). Besides the presence of elevated levels of advanced glycation endproducts (AGEs) as a result of high glucose concentrations [3,62,63], they hypothesized a deficit in IVD nutrition caused by changes in the thickness and porosity of the bony vertebral EP and, subsequently, reduced microvascularity and vessel diameters in response to T2DM (Figure 3, [53]) as a trigger of IVDD development. Malnutrition also leads to enhanced expression of hypoxia-inducible genes in the IVD [15] (Figure 4).  Table 4. Molecular events in the IVD are depicted in Figure 4. AF: annulus fibrosus, BEP: bony endplate, CEP: cartilaginous endplate, T2DM: type 2 diabetes mellitus, NP: nucleus pulposus, IVD: intervertebral disc, ↓: decrease, ↑: increase.
The high levels of AGEs mediate an intensified interaction with the Receptor for advanced glycation end-products RAGE receptor [69]. AGEs represent pathogenetic risk factors. They are increased in T2DM [15], but they are also risky components of modern diets, and hence, the effect of chronic consumption of diets with high levels of AGE on IVDD was addressed in several of the  Table 4. Molecular events in the IVD are depicted in Figure 4. AF: annulus fibrosus, BEP: bony endplate, CEP: cartilaginous endplate, T2DM: type 2 diabetes mellitus, NP: nucleus pulposus, IVD: intervertebral disc, ↓: decrease, ↑: increase.
The high levels of AGEs mediate an intensified interaction with the Receptor for advanced glycation end-products RAGE receptor [69]. AGEs represent pathogenetic risk factors. They are increased in T2DM [15], but they are also risky components of modern diets, and hence, the effect of chronic consumption of diets with high levels of AGE on IVDD was addressed in several of the selected studies [3,52,62,63]. The AGE/RAGE interaction results in amplified autophagy detectable by the expression of diverse autophagy markers in IVD cells [56]. Autophagy is a strategy of stressed cells to eliminate damaged cell organelles. In addition, IVD and NP cells undergo senescence [55,57]. Moreover, mitochondrial stress, characterized by reactive oxygen species (ROS) accumulation but also associated with compensatory elevation of antioxidants, such as catalase and manganese superoxide dismutase (MnSOD), could be detected [54]. As an additional feature of mitochondrial stress, the mitochondrial membrane potential is disrupted [59,67], and a disbalance of mitochondrial B-cell lymphoma 2 (Bcl-2) family apoptosis regulators could be observed [67]. Elevated pro-apoptotic Bax induces the release of cytochrome c from the mitochondrial intermembrane space into the cytosol, and thus, the apoptosis cascade proceeds [74]. Accordingly, increasing the activity of initiator and executioner caspases is observed [67].
IVD cell stress under high glucose conditions is also characterized by the increased release of inflammatory cytokines, such as Tumor Necrosis Factor (TNF)α [61,68], interleukin (IL)-1, and IL-6 [61]. These cytokines contribute to the unrestrained induction of diverse matrix-metalloproteinases (MMPs) and A disintegrin and metalloproteinase with thrombospondin motifs (ADAM-TS), which degrade a wide range of ECM components, leading to the loss of IVD ECM integrity and stability [61,62]. In addition to cell loss by apoptosis, the specific anabolic synthesis profile of IVD cells is impaired by cell stress and senescence in response to high glucose, which further weakens the IVD ECM and, therefore, accelerates its degeneration. With regard to this, the anabolic growth factors-IGF-1 and TGFβ-are reduced, which are known to stimulate ECM synthesis [68]. Thus, IGF-1 treatment is proposed as a strategy to inhibit IVD degeneration under T2DM conditions [47]. The expression of IVD-associated ECM proteins, such as collagen type II, I, GAGs, aggrecan, and the chondrogenic transcription factor SOX9, is suppressed under high glucose [15,62,64]. Moreover, ECM composition might also change due to IVD cell trans-differentiation since hypertrophic markers, such as collagen type X, could be detected in IVD cells in response to hyperglycemia [52]. Another study observed a loss of tissue organization, such as IVD fibrosis, and the absence of the AF lamellar pattern, suggesting that the order of collagen fiber bundles disappeared under hyperglycemia [61]. Therefore, the capability of the IVD to distribute homogeneously the compressive strain might be impaired under these conditions.
Because abundant AGEs associated with T2DM also enhance collagen cross-linking [3,27] responsible for increased IVD stiffness and loss of GAGs and proteoglycans, which impair water binding and viscoelasticity, the overall biomechanical properties and adaptability of IVD are severely altered under T2DM conditions. As a final key feature, this impaired IVD biomechanics ultimately leads to IVDD with increasing the risk of herniation and resulting in radiculopathies [69]. Apart from suggesting the involvement of activated mitogen-activated protein (MAP) kinases (extracellular signal-regulated kinase: ERK and p38) in the interrelation of T2DM and IVDD [58,60,63], signaling pathways shared by T2DM and IVDD pathogeneses were further addressed in detail in the 23 studies included. Only two studies investigated the more specified pathogenesis by addressing the role of Sirt1/p53 interplay in the interrelation between DM and IVDD in regard to NP cell apoptosis [58]. They evaluated the role and the underlying mechanism of MALAT1 in the apoptosis of cartilage endplate (CEP) cells induced by high glucose concentrations [58] providing more detailed analysis of the mode of interaction.
In regard to the pathogenesis of IVDD under T2DM conditions, the resulting hyperglycemia obviously affects the whole motion segment, embracing the vertebrae [52,62,70] and the components of the IVD (AF and NP) between them [43,61]. Despite the present study focused on the IVD, the hyaline cartilaginous part of the EP and vascularization of the bony EP were addressed because both represent an important precondition for IVD nutrition [53].
Two studies presented also putative therapeutical options by treatment with glucose suppressing injection of IGF-1 [47] or with butein as an agonist of protective Sirt1 [58]. Two studies addressed options for IVDD prevention under high glucose conditions. The authors of the first study tested the effect of vitamin D (calcitriol) on IVDD in DM, which could partially restore TGFβ and IGF-1 in the discs [68]. The other research group found that the oral treatment with a combination of anti-inflammatory and anti-AGE drugs (pentosan-polysulfate and pyridoxamine) was able to reduce to some degree diabetes-induced degenerative changes in the vertebra and IVD [62].  Table 4. ADAMTS: a disintegrin and metalloproteinase with thrombospondin motifs, AGE: advanced glycosylation endproducts, Bax: Bcl-2 associated protein X, Bcl-2: B-cell lymphoma 2, ECM: extracellular matrix, GAG: glycosaminoglycans, IGF-1: insulin growth factor-1, IL: interleukin, IVDD: intervertebral disc degeneration, MMP: matrix metalloproteinase, RAGE: a receptor for advanced glycation endproducts, ROS: reactive oxygen species, SIRT: sirtuin, pRB: retinoblastoma protein, SOX9: sexdetermining region SRY of the Y chromosome, TIMPs: tissue-specific inhibitor of matrix metalloproteinases, TNFα: tumor necrosis factor α, T2DM: diabetes mellitus type II, TGFβ: transforming growth factor β.
In regard to the pathogenesis of IVDD under T2DM conditions, the resulting hyperglycemia obviously affects the whole motion segment, embracing the vertebrae [52,62,70] and the components of the IVD (AF and NP) between them [43,61]. Despite the present study focused on the IVD, the hyaline cartilaginous part of the EP and vascularization of the bony EP were addressed because both represent an important precondition for IVD nutrition [53].
Two studies presented also putative therapeutical options by treatment with glucose suppressing injection of IGF-1 [47] or with butein as an agonist of protective Sirt1 [58]. Two studies addressed options for IVDD prevention under high glucose conditions. The authors of the first study tested the effect of vitamin D (calcitriol) on IVDD in DM, which could partially restore TGFβ and IGF-1 in the discs [68]. The other research group found that the oral treatment with a combination of anti-inflammatory and anti-AGE drugs (pentosan-polysulfate and pyridoxamine) was able to reduce to some degree diabetes-induced degenerative changes in the vertebra and IVD [62].
Using animal models, sex aspects have to be accounted for. Most of the research groups (Table  4) selected male rodents for their experiments. Unfortunately, some of the summarized studies did not provide any sex information [15,43,52,68]. Sex differences are barely investigated. Mostly male animals were used (18 of 23 studies, four studies provided no sex information). Two research teams compared males and females and found that AGE in diet had more effects in females [3,70], and one study used only female mice [62]. Krishnamoorthy and Natelson et al. compared the effect in males with female mice, elucidating a clear association with more severe effects on IVD biomechanics in females [3] and impaired bone structure [70]. In contrast, regarding a leptin receptor-resistant model, Li et al. stated a protective effect in females due to sex steroids [47] as a justification to include only males in their experimental groups. The observable sex effects might depend on the model (mouse vs. rat, different lines of both species, specific knockout) or also on the mode of T2DM induction used and should necessarily be investigated in more detail in the future.  Table 4. ADAMTS: a disintegrin and metalloproteinase with thrombospondin motifs, AGE: advanced glycosylation end-products, Bax: Bcl-2 associated protein X, Bcl-2: B-cell lymphoma 2, ECM: extracellular matrix, GAG: glycosaminoglycans, IGF-1: insulin growth factor-1, IL: interleukin, IVDD: intervertebral disc degeneration, MMP: matrix metalloproteinase, RAGE: a receptor for advanced glycation end-products, ROS: reactive oxygen species, SIRT: sirtuin, pRB: retinoblastoma protein, SOX9: sex-determining region SRY of the Y chromosome, TIMPs: tissue-specific inhibitor of matrix metalloproteinases, TNFα: tumor necrosis factor α, T2DM: diabetes mellitus type II, TGFβ: transforming growth factor β.
Using animal models, sex aspects have to be accounted for. Most of the research groups (Table 4) selected male rodents for their experiments. Unfortunately, some of the summarized studies did not provide any sex information [15,43,52,68]. Sex differences are barely investigated. Mostly male animals were used (18 of 23 studies, four studies provided no sex information). Two research teams compared males and females and found that AGE in diet had more effects in females [3,70], and one study used only female mice [62]. Krishnamoorthy and Natelson et al. compared the effect in males with female mice, elucidating a clear association with more severe effects on IVD biomechanics in females [3] and impaired bone structure [70]. In contrast, regarding a leptin receptor-resistant model, Li et al. stated a protective effect in females due to sex steroids [47] as a justification to include only males in their experimental groups. The observable sex effects might depend on the model (mouse vs. rat, different lines of both species, specific knockout) or also on the mode of T2DM induction used and should necessarily be investigated in more detail in the future.
In regard to the type of rodent in vivo models, rat models dominated (eight of 13 in vivo studies) and were often based on SD rats treated with streptozotocin (STZ) [53,58,63,64,68]. The preference of rats the fact that all in vitro studies were based on rat-derived cells might be due to the larger sample size in rats compared to mice. The preferential use of STZ-treated SD rats compared to genetically-modified rats might be influenced by costs.
Bearing in mind that the spine function depends on its unique biomechanics and T2DM might contribute to biomechanical failure; unfortunately, biomechanical investigations were undertaken by just a few of the selected in vivo studies. Looking at the in vitro studies, all of them used monolayer cultures, but no three-dimensional (3D) culture models were included, which might mimic more closely the in vivo conditions. 3D culture could also allow biomechanical testing in vitro in future DM cell models.
However, this literature review revealed four strong, 17 moderate, and one low evidences of the link between DM and IVDD in rats in vivo and in vitro, which clearly underlines the effect of DM on the development of IVDD, while only one study reported an absence of the evidence [70].
The number of rodents used varied greatly in the in vivo studies, usually ranging between 20 and 48 individuals. One study [75] used a huge number of experimental animals (about 255 rats) divided into two groups. The in vitro studies generally used fewer individuals for cell harvesting. In these studies, performed in vitro, generally, only one group of non-diabetic rats (mostly SD rats) gave the needed IVD cells, which were then incubated in the culture medium of two different glucose concentrations as experimental and control groups. The direct comparison to human samples or IVD cell population and estimation of transferability of results remains mostly unclear (one has to be aware that there are notochordal cells in rodents. Furthermore, in rodent models, often coccygeal IVDs are used because they are larger than lumbar discs in the rodents [15]). Only the in vitro study of Tsai et al. included parallel experiments with human cells extracted from IVD tissues of healthy and diabetic patients, suggesting a nice parallel agreement of the results [63].
Worth to mention is that some crucial approaches, such as biomechanical investigations, are underrepresented in the included pool of studies.

Study Strength
The strength of this study is based on the large number of the included studies, which linked in vivo and in vitro results. In addition, it simultaneously summarized studies, which analyzed the pathogenesis, treatment, prevention, and risk factors of the issue of interest. Altogether, this presents a plethora and wide range of detected positive findings. Moreover, it covers a long period of data generation extending from 1992 to 2020.

Study Limitations
One limitation of our study may be the literature-based study design, whereby the group sizes, the particular type of rodent model, focus, and down-stream analyses of the studies summarized and discussed here differed from each other. It might also hamper the fact that in vivo and in vitro studies were put together to get a sufficient number of relevant studies included.
In addition, mouse and rat models were summarized, which might bear some heterogeneity.

Conclusions
T2DM and IVDD are likewise diseases with a high prevalence, incidence and distribution, affecting individuals with different socioeconomic levels. Therefore, both have a high public health burden worldwide. Many diseases have exhibited themselves as complications or co-morbidities in diabetic patients. The evidence of the role of DM in the development of IVDD based on rodent models was clearly detected ranging between strong and moderate in power. Hence, based on the collective findings from the experiments performed in vivo and in vitro, our hypothesis regarding the link between T2DM and IVDD is strongly supported. However, collecting a variety of aspects from the available literature that contribute to the linked pathogenesis (Figure 4), the shared signaling pathways behind it are still hidden and need to be explored in the future in order to identify effective therapeutical targets.