Characterisation of osteoarthritis in a small animal model of type 2 diabetes mellitus

Article focus

It has long been acknowledged that type 2 diabetes mellitus (T2DM) is a probable factor in the onset and progression of osteoarthritis. However, there is yet to be an appropriate small animal model for T2DM due to the nature of the disease. Without one, it is difficult to understand the mechanisms behind its action and address its effects.

Our study’s aim was to characterise the quality of the articular cartilage at the level of the knee in T2DM rats, thus determining the model’s effectiveness in the study of osteoarthritis.

Key messages

T2DM contributed to osteoarthritis progression in several areas of the rat knee joint.

The animal model used seems well suited for further investigation of the effects and mechanisms of T2DM, at least in the development of osteoarthritis.

Strengths and limitations

Strength: The interpreters were blinded to the conditions when scoring the histology

Limitations: We had a limited number of T2DM rats as they were bred by the UCD group. Their growth and diet were not monitored by our group.

The grading scale used is subjective.

The rats used in this study were not age-matched, and the age among rats within the T2DM group also varied.

Introduction

Type 2 diabetes mellitus (T2DM) is a growing global epidemic, currently affecting over 200 million people.¹ T2DM is characterised by reduced pancreatic beta-cell function and systemic insulin resistance, leading to metabolic dysfunction throughout the body. The exact pathophysiological mechanism behind T2DM remains largely unknown, however, the disease is associated with obesity, a high-fat diet, and a lack of physical activity.^2-5 In the field of orthopaedic surgery, T2DM has mostly been studied as a surgical comorbidity.⁶ However, there are aspects of T2DM that are of interest in musculoskeletal research, particularly articular cartilage biology.

Factors such as diet, increased body mass and obesity have been studied in both clinical and basic scientific research, as they relate to osteoarthritis onset and progression.^3,7,8,9-15 Additionally, the literature also investigated the role that increased adipose tissue plays in increased inflammatory response in the joint tissues.^7,16,17 For instance, there is a positive correlation between elevated blood glucose levels and evidence of OA.^10,11,14,18 These conditions seem to provide catabolic signals that lead to increased proteolytic enzyme activity that degrades the matrix components, causing progressive cartilage degradation.¹⁸ Proteolytic enzymes belonging to the metalloproteinase family, such as collagenases and aggrecanases, are especially relevant as their expression is increased by pro-inflammatory and catabolic stimuli, and whose activity has been shown to be augmented in OA cartilage.¹⁸ Studies have also attributed the development of osteoarthritis to metabolic disorders and to the stiffening of tissues through an accelerated form of non-enzymatic cross-linking of collagen often associated with ageing.¹⁹ Briefly, increased circulating glucose reacts with free amino acids on proteins through the Maillard reaction, a form of non-enzymatic browning resulting from a chemical reaction of an amino acid and a reducing sugar, to create stable intermediates. These glycated protein intermediates then link together to form larger, insoluble aggregates known as advanced glycation end products (AGEs).²⁰ This phenomenon occurs in collagen-rich tissues, such as cardiovasculature and bone throughout the body, and leads to mechanical and biochemical changes in those tissues.^20-23 AGEs naturally accumulate through the process of ageing and have been studied as a possible mechanism for spontaneous, or primary, osteoarthritis in aged populations.²²

One of the major hurdles for in vivo T2DM study has been the lack of an appropriate small animal model. Although there are several different transgenic, surgical, and spontaneous rodent strains that have been used to model T2DM, the diabetic physiology of these animals deviates significantly from that seen in humans.^19,24 One example of this difference is the fact that symptoms of T2DM in rats, such as obesity and insulin resistance, result largely from monogenic mutations rarely found in human populations.⁷

In 2008, Cummings et al²⁴ characterised the Graham-Havel T2DM strain of rats created by their group at the University of California, Davis (UCD). Crossing obese, insulin-resistant Sprague-Dawley rats with Zucker diabetic fatty-lean rats, which do not develop obesity or insulin resistance, was the first step in creating these rats. The progeny of these animals are then selectively bred to the F7 generation, at which point they exhibit the T2DM phenotype.²⁴ This resulted in the UCD-T2DM rats with a polygenic phenotype similar to that seen in humans.²⁴ Furthermore, this strain does not exhibit infertility that is characteristic of other experimental diabetic rodent lines, making them hard to maintain.^19,24 In these animals, the onset of T2DM was confirmed using fasting glucose levels, glucose tolerance, immunohistochemistry of the pancreas, and islet insulin content in both sexes.²⁴

An appropriate small animal T2DM model will allow not only a better understanding of the onset of T2DM, but also of the mechanisms by which T2DM contributes to a host of diseases and disorders. The purpose of our study is to characterise the quality of the articular cartilage at the knee in the UCD-T2DM rats. We hypothesise that this strain of rats will have more severe osteoarthritis in comparison with control rats.

Materials and Methods

Results

Safranin-O staining

These scores are based on the grading scale in Table I. Control joints had a mean (standard deviation, sd) osteoarthritis score of 1.59 (sd 0.438) at the lateral tibia, 1.16 (sd 0.475) at the lateral femur, 1.22 (sd 0.466) at the medial tibia, and 0.50 (sd 0.231) at the medial femur. The UCD-T2DM had mean osteoarthritis scores of 2.43 (sd 0.229) at the lateral tibia, 2.81 (sd 0.311) at the lateral femur, 2.84 (sd 0.718) at the medial tibia, and 1.68 (sd 0.213) at the medial femur. Therefore, the UCD-T2DM rat had a significantly higher, thus worse, osteoarthritis score at the lateral femur (p = 0.013), medial tibia (p = 0.003), and medial femur (p = 0.034). Control and UCD-T2DM joints did not differ significantly at the lateral tibia (p = 0.158). Figure 1 shows the difference in osteoarthritis scores between the control and the UCD-T2DM rat knees.

Fig. 1

Graph comparing osteoarthritis scores between the control and the UCD-T2DM rat knees. The scoring scale used is modified from Pritzker et al²⁵ “*” denotes statistically significant differences of p < 0.05. The UCD-T2DM rat had a significantly higher, thus worse, osteoarthritis score at the lateral femur (p = 0.013), medial tibia (p = 0.003), and medial femur (p = 0.034) (Wilcoxon Test).

The lateral and medial femoral condyles show both degeneration of the articular surface and areas of staining loss relative to the surrounding tissue. The articular cartilage was also phenotypically different at the medial condyles between the two groups. Control rats showed normal healthy articular cartilage at this surface, while the UCD-T2DM rats displayed cysts within the articular cartilage and at the junction between the articular cartilage and subchondral bone. Figures 2 and 3 show Safranin-O stained knees for both normal and UCD-T2DM rats in the lateral and medial aspects, respectively; arrows show areas of injury phenotype.

Fig. 2

Micrographs showing the lateral aspect; representative sections stained with Safranin-O/Fast Green, taken at 5x magnification. Arrows indicate areas in which erosion of articular cartilage, loss of staining, and other irregularities have occurred in the UCD-T2DM group. On the lateral aspect of the knee in the UCD-T2DM group, there is prominent reduction in articular cartilage quality at the femoral condyle and thinning of the cartilage at the tibial plateau.

Fig. 3

Micrographs showing the medial aspect; representative sections stained with Safranin-O/Fast Green, taken at 5x magnification. Arrows indicate areas in which erosion of articular cartilage, loss of staining, and other irregularities have occurred in the UCD-T2DM group. On the medial side, the UCD-T2DM group showed cysts within the cartilage and pockets of chondrocyte loss along the tibial articular surface.

Immunohistochemistry

There does not appear to be an increase in type II collagen expression between the two rat groups. Type II collagen stained strongly at the chondrocytes throughout the articular cartilage. It was observed that the number of articular chondrocytes in the UCD-T2DM knees appears to be less than in normal knees, particularly in the lateral femoral condyle (Fig. 4). Type I collagen expression at the articular cartilage in the UCD-T2DM rats appears to remain limited to the superficial surface, in a similar way to the control rats. Figure 4 shows representative micrographs at 63x magnification of sections stained for col1a1 and col2a1. Figure 5 likewise shows representative micrographs of the medial femoral condyle.

Fig. 4

Micrographs showing the lateral aspect; series of representative sections stained with IHC for col I and col II, taken at 63x magnification.

Fig. 5

Micrographs showing the medial aspect; series of representative sections stained with IHC for col I and col II, taken at 63x magnification.

Mass and correlation with osteoarthritis

The masses of the rats used in this study were recorded before dissection. Control Lewis rats had a mean mass of 266 gm (sd 4.72). The UCD-T2DM rats had a mean mass of 592 gm (sd 16.0). A Welch Two-Sample t-test was used to compare the two rat groups, showing that the mass of these specimens is significantly different by strain (p < 0.001).

To explore the relationship between the mass of rats and their respective osteoarthritis score, a one-way ANOVA was conducted at each of the different regions of the knee. There was a statistically significant correlation between mass and histology score at the lateral femur (p = 0.011), medial femur (p = 0.003), and medial tibia (p = 0.031). The lateral tibia was not significantly correlated at p = 0.051, however, it can still be considered a strong relationship.

Figure 6 summarises the differences between the rats with respect to mass, while Figure 7 shows the correlation between mass and osteoarthritis severity by region.

Fig. 6

Bar chart showing the comparison of mass between control and UCD-T2DM rats. “*” denotes statistically significant difference in weight between the UC Davis T2DM rats and the control Lewis rats (p < 0.001) (Welch Two-Sample t-Test).

Fig. 7

Graphs showing the correlation between osteoarthritis score and mass at different regions of the knee: a) lateral femur, b) lateral tibia, c) medial femur, and d) medial tibia. There is significant correlation between mass and histological score at the lateral femur (p = 0.011), medial femur (p = 0.003), and medial tibia (p = 0.031). The lateral tibia is not significant, but can still be considered a strong correlation (p = 0.051) (one-way ANOVA).

The mean age of the UCD-T2DM rats used in this study was 228 days (sd 19.4). This significantly differed from the six-month-old (180 days) controls used in this study (p = 0.032). This may be due to the necessary length of time for the UCD rats to develop osteoarthritis before sacrifice. When considering the relationship between age and severity, there was no correlation between osteoarthritis score and T2DM at either femoral condyle or the lateral tibial plateau. However, at the medial tibial plateau there was statistically significant correlation with osteoarthritis severity (p = 0.046).

Discussion

The aims of this study were to histologically characterise the knees in rats with a unique type 2 diabetes phenotype and compare them with the knees in young healthy rats. Histologically, we were able to demonstrate that there was mild, global degeneration of the articular surfaces of the knee. Our immunohistochemistry stains did not show a change from type II collagen, normally found in the articular joint, to type I collagen, Overall, the results from this study suggest that this animal model can be used to further study the aspects of T2DM that contribute to the development and progression of osteoarthritis.

Type 2 diabetes is a fast-growing epidemic throughout the world, both in ‘first-world’ and developing countries.²⁹ T2DM has been associated with a variety of musculoskeletal disorders, including osteoporosis and osteoarthritis.^2,4-6,30 A 20-year longitudinal and cross-sectional study provided evidence that T2DM should be considered a risk factor independent of other factors such as body mass.⁴ However, the pathophysiological relationship between T2DM and osteoarthritis remains largely unknown.

The animal model used in this study, the Graham-Havel UC-Davis type II diabetic rat, has a polygenic phenotype similar to that of humans.²⁴ Animals used in previous studies tended to be monogenic and had comorbidities such as infertility, not associated with human type 2 diabetes.^23,24 The UCD type 2 diabetic rats not only had reduced insulin sensitivity, but also had significantly more body mass.²⁴ This study clearly demonstrates that in the absence of an external injury, the UCD-T2DM had a lower quality of cartilage across regions of the knee, compared with control rats.

Body mass appeared to be an important factor based on the correlation with osteoarthritis score in this study. More severe phenotypes at the lateral femur, medial tibia, and medial femur were significantly correlated with greater mass (p = 0.011, 031 and 0.003, respectively). The lateral tibia, although not significantly different (p = 0.051), appeared to have a strong trend based on the one-way ANOVA. However, increased mass could potentially be a confounding factor in this study. It was formally believed that simply increasing the load to the knee, a weight-bearing joint, by increasing body mass, was enough to induce osteoarthritis. However, several studies have called the simplicity of that mechanism into question.^4,9 It was suggested by Felson et al,⁹ based on a study of human radiographs, that obesity is correlated with osteoarthritis in the knee when a joint is misaligned. Furthermore, Mooney et al¹⁴ demonstrated that post-traumatic osteoarthritis progression was accelerated by a high-fat diet, but that this phenomenon was independent of weight gain. Future studies should include a body mass control to determine whether there are histological differences between groups having T2DM and having obesity. Studying joints bearing less weight than the knee, such as the shoulder, may also be beneficial in determining whether T2DM has effects independent of osteoarthritis progression.

Studies have suggested cross-linking of collagen fibres and the accumulation of aggregated glycation end-products (AGEs) as a potential mechanism for progression of osteoarthritis by ageing.²²In vitro, non-enzymatic cross-linking of collagen fibres in cartilage tissue was achieved by incubating samples in media containing threose, ribose or methylglyoxal.^22,23,26 Monnier et al²⁰ showed that T2DM accelerated the accumulation of AGEs in skin samples from diabetic subjects.

Synthesising these studies together, T2DM may represent a type of advanced ageing, with more rapid accumulation of AGE than with age alone. Furthermore, Wagner et al²³ demonstrated that human spinal disc tissue treated with methylglyoxal contained significantly increased amounts of AGEs, changing the biomechanical properties of the tissue. The implication here is that the stiffening of the collagen network changes the biomechanical properties of the tissues, making them less able to deform and potentially more susceptible to injury.

However, as pointed out by Willet et al,³¹ these in vitro studies have yet to show a direct causative effect between AGEs and the progression of osteoarthritis. In vivo studies by this group and Vos et al³² have shown that despite significantly increasing the amount of pentosidine, one form of AGE, osteoarthritis progression is not accelerated. Vos et al³² surgically induced canine osteoarthritis and were not able to show an effect of AGE on osteoarthritis severity at the time of onset by intra-articular treatment of ribose and threose. Willet et al³¹ also showed that artificially induced cross-linking with intra-artciular injections of 2M ribose in a spontaneous osteoarthritis model in Hartley guinea pigs, did not change the natural history of the disease compared with animals injected with saline only. This group quantified the non-enzymatic glycation of the articular cartilage, claiming they had achieved significantly increased cross-linking, comparable with the level typical in an 80-year-old human.³¹ Both models show that the cross-linking did not create a different phenotype from the injury controls. Thus, despite the evidence from in vitro models, the more recent in vivo models would suggest that cross-linking of collagen alone without the metabolic effects of T2DM may not be sufficient for the induction of osteoarthritis.

There are other indirect effects of cross-linking and obesity that could contribute to the understanding of T2DM-associated osteoarthritis. First, obesity, or more specifically the increase of adipose tissue, leads to an increase in circulating adipokines. Clinically, these adipokines have been shown to have a correlation with the degeneration of the articular cartilage.^16,17 De Boer et al¹⁷ not only showed a relationship between upregulation of adipokines and osteoarthritis radiologically, but also confirmed these results histologically and biochemically. Aditionally, this group presented evidence to suggest that adipokines such as lectin and adiponectin also act on the synovial tissues of the joint, increasing the presence of pro-inflammatory cytokines such as interleukin-1β (IL-1β) and tumour necrosis factor-α (TNF-α).¹⁷

Although we have shown that there is a clear and significant histological difference between healthy rats and the UCD-T2DM rats, our study does possess some limitations that require us to temper our conclusions. Because the UCD-T2DM rats were frozen and then thawed prior to fixation, it is possible that morphology was affected. However, cryotechniques are commonly used in other histological studies, and have been able to maintain cellular morphology in vitro even at freezing; as a result, the morphology in the UCD rats should also remain unchanged.³³

Genders were not tracked in the UCD-T2DM group, which may introduce some bias, although the population should have been evenly distributed. The rats used in this study were not age-matched, and the age among rats within the T2DM group also varied due to the differing onset of diabetes. However, the variance in ages did not have a significant impact on osteoarthritis severity except in the medial tibia, thus it was not considered a significant confounding variable.

The marked mass difference between the T2DM rats and control could also provide a bias in the data. It is also difficult to extrapolate the aspects of T2DM that contribute to the osteoarthritic phenotype, and we can only speculate based on the other studies presented here. Furthermore, we did not have enough animals in this study to harvest enough cartilage for biochemistry. However, this is beyond the scope of this study and should be considered for future work; the main objective was to confirm that the UCD-T2DM rats had an osteoarthritic phenotype at their knees.

This work shows a high prevalence of osteoarthritis in the UCD-T2DM rat knee. These rats differ from the control animals based upon their weight and presence of T2DM. We believe that both obesity and the advanced accumulation of the AGEs may be contributing to this osteoarthritic phenotype, however, the manner of this still remains unclear. The strength of this animal model is in the creation of a metabolic syndrome environment in these rats. Interventions that would be effective in clinical practice will also be effective in setting the multifactorial causation of T2DM. We suggest that further studies with these rats can focus on the control of weight and/or T2DM on the progression of osteoarthritis.