Subchondral bone remodeling patterns in larger animal models of meniscal injuries inducing knee osteoarthritis – a systematic review

Purpose Elucidating subchondral bone remodeling in preclinical models of traumatic meniscus injury may address clinically relevant questions about determinants of knee osteoarthritis (OA). Methods Studies on subchondral bone remodeling in larger animal models applying meniscal injuries as standardizing entity were systematically analyzed. Of the identified 5367 papers reporting total or partial meniscectomy, meniscal transection or destabilization, 0.4% (in guinea pigs, rabbits, dogs, minipigs, sheep) remained eligible. Results Only early or mid-term time points were available. Larger joint sizes allow reporting higher topographical details. The most frequently reported parameters were BV/TV (61%), BMD (41%), osteophytes (41%) and subchondral bone plate thickness (39%). Subchondral bone plate microstructure is not comprehensively, subarticular spongiosa microstructure is well characterized. The subarticular spongiosa is altered shortly before the subchondral bone plate. These early changes involve degradation of subarticular trabecular elements, reduction of their number, loss of bone volume and reduced mineralization. Soon thereafter, the previously normal subchondral bone plate becomes thicker. Its porosity first increases, then decreases. Conclusion The specific human topographical pattern of a thinner subchondral bone plate in the region below both menisci is present solely in the larger species (partly in rabbits), but absent in rodents, an important fact to consider when designing animal studies examining subchondral consequences of meniscus damage. Large animal models are capable of providing high topographical detail, suggesting that they may represent suitable study systems reflecting the clinical complexities. For advanced OA, significant gaps of knowledge exist. Future investigations assessing the subchondral bone in a standardized fashion are warranted.


Introduction
Understanding the spatio-temporal trajectory of subchondral bone remodeling may provide better insights into osteoarthritis (OA) [27].In the knee, both menisci complement and protect the osteochondral unit [25].Meniscus injury is common [20,32], and meniscus tissue insufficiency and loss are one of the most important causes of knee OA [5-7, 21, 49, 52].The principle of inducing a defined meniscal lesion is applied in many animal models to model OA initiation and development [42,50].However, the resulting subchondral bone alterations are incompletely understood [36], in contrast to the cartilage [35].Though rodents represent by far the most popular model, their small joints limit an accurate topographical recapitulation.Therefore, this systematic review focuses on larger animal models only, analyzing the available knowledge about subchondral bone remodeling applying traumatic meniscal injuries as a standardizing entity.It reports study designs, establishes missing gaps and pinpoints future research directions.

Comparative morphology of the subchondral bone
As its two major parts, the subchondral bone plate and subarticular spongiosa are structurally dissimilar, they need to be considered separately [37].The human subchondral bone plate is composed of 0.2-0.4mm thick plates joining together, enclosing pores, extensions of the marrow space and invading vascular channels [16].Humans have the largest tibial plateau width (~ 7.4 ± 0.5 cm), followed by sheep (~ 5.1 ± 0.1 cm), minipigs (~ 3.9 ± 0.1 cm), rabbits (~ 1.6 ± 0.1 cm), rats (~ 0.7 ± 0.1 cm), and mice (~ 0.3 ± 0.1 cm).Remarkably, the human subchondral bone plate is considerably thinner (0.52 ± 0.11 mm) than in sheep (1.32 ± 0.14 mm), minipigs (0.82 ± 0.17 mm) and similar to rabbits (0.49 ± 0.05 mm) [39].In these animals, the subchondral bone plate is more compact and less porous [39].When normalized to the tibial plateau width, the subchondral bone plate in minipig is twofold, and in sheep and rabbits threefold thicker than in humans [39].Its microstructure (Table 1) relates to the severity of OA [24,26,34,40].The subchondral bone plate of rabbits and minipigs is most similar to humans, including similar BS/BV, BS/TV, and closed porosity, while BV/TV is higher, and open and total porosity are lower [39].In sheep, BV/TV is higher, and BS/BV, BS/ TV, open and total porosity are lower than in humans [39].Of note, the specific human topographical pattern of a thinner subchondral bone plate in the region below both menisci is present solely in the larger species (partly in rabbits), but absent in rodents, a clinically highly relevant fact to consider when designing animal studies examining structural (subchondral) consequences of meniscus damage.

Pattern of subchondral bone changes
Structural subchondral bone alterations are important characteristics of both early and advanced stages of OA.At onset, primary osteoporotic changes occur [10], besides early degenerative cartilage changes [15,25,36,43,45].Subchondral bone plate porosity increases, and BMD, trabecular volume, and complexity of the trabecular structure decrease [15,43].These changes may be caused by microdamage of the trabeculae due to altered load [17,28].
After the initial bone loss, still in early OA, subchondral sclerosis (increase in trabecular volume and complexity), as well as osteophytes occur [25,43].Later, pronounced abnormalities of bone shape and cysts appear [25].

(a) PubMed search
An initial PubMed search was performed on 15.04.2023 with the terms "(osteoarthritis) AND ((meniscus) OR (meniscal) OR (meniscectomy))" yielded 5367 results, including several studies not reporting any subchondral bone data (Fig. 1).When the search was refined as "(subchondral bone) AND (osteoarthritis) AND ((meniscus) OR (meniscal) OR (meniscectomy))", it resulted in 521 papers (9.7% of the original search).In the detailed analysis only those studies were included where OA was evoked by traumatic tear or injury of the meniscus, surgical destabilization of the meniscus (DMM), or total or partial meniscectomy.To avoid any supplementary factors leading to additional instability such as ligament transection (e.g.anterior cruciate ligament), such combined methods were excluded (Fig. 1).Papers were also excluded if full text was unavailable, if they were reviews, not in English language, subchondral bone or appropriate controls not reported, or meniscal injury not applied to induce OA.The paper selection consequently was further reduced after reading their abstracts (n = 222), and full text (n = 152 papers).When human (n = 9), mouse (n = 96) and rat (n = 25) studies were excluded, n = 23 studies remained eligible for detailed systematic evaluation (only 0.4% of all papers reporting meniscal injury).

(b) Evaluation of the abundance of studies
The first animal study fulfilling the inclusion criteria was published in 1993 [3].The abundance of eligible papers was constantly low afterwards (n = 0-6 in 4-year intervals) (Fig. 2a).

(c) Evaluation of the study designs
In the finally selected 23 papers, four main types of induced meniscal damage were described: (1) total meniscectomy, (2) partial meniscectomy, (3) meniscal transection or tear (MMT; the medial collateral ligament [MCL] is transected in small animals and left intact in large ones, and the pars intermedia is transected at its narrowest point, but no parts of the meniscus are removed), and (4) DMM by "meniscal release" (the anterior root is transected, but no parts are removed).

(d) Evaluation of study methods
The most common method to evaluate subchondral bone structure was histology, used in 44% of the studies (n = 10).Micro-CT was used only in 39% of all studies (n = 9), although it represents the gold-standard method for directly evaluating bone microstructure [9], with excellent reproducibility and accuracy [8] (Fig. 2b).When the occurrence of applying a combined evaluation protocol including microcomputed-tomography (micro-CT) or histology, paired with dual energy X-ray absorptiometry (DXA), biochemistry or gross pathology of the joint was examined, histological evaluation was mostly used alone (40% of the total n = 23 studies), or in combination with gross pathology (30%), or DXA (30%).Micro-CT was used alone in 30% of all studies, and in combination with gross pathology in 9%.Histological evaluation was mostly applied for reporting subchondral bone plate thickness, but several studies also used it to evaluate subchondral trabecular microstructure, despite the strong limitation of the stereologic analysis of a few 2-dimensional (2D) sections assuming plate-like underlying structure [8].Many of such studies did not identify significant differences between treatment groups [23,33].Overall, the most frequently reported bone parameters were BV/TV, BMD, presence of osteophytes, Tb.Th, thickness of the subchondral bone plate, and Tb.N (Fig. 2c).Larger animal studies only reported early and mid-term time points (Fig. 2d).
In sum, meniscal damage at 3 months in guinea pigs resulted in increased subchondral bone plate thickness and loss of trabecular bone (Table 2, Fig. 3a-c), similarly to other early/mid-term OA models.
In sum, total meniscectomy induced only minor changes in the subchondral bone, including decreased subchondral bone plate BMD and increased bone blood flow at the earlier time points (2-8 weeks).Osteophytes developed in the later phase after 12 weeks, while BMD and trabecular structure were unchanged.
In sum, DMM induced only minor changes of the subchondral bone after 8-12 weeks, including increased BV/ TV, and trabecular bone loss mostly in the compartment opposing the operated compartment.The fact that no characteristic major microstructural changes were detected (Fig. 3d, e) might be due to the limited sensitivity of the applied 2D detection methods applied.In order to achieve better comparability with human and other animal data, more short-and long-term studies with sensitive 3D detection methods such as micro-CT are needed.
Thus, data from dogs are scarce and analyses only based on 2D histological sections which have limited accuracy compared to true 3D analyses with micro-CT [8].Furthermore, time points covered only the 12-16 week period corresponding to early/mid stage OA, when no extensive alterations of the subchondral bone were observed.To allow for a detailed comparison with human and other animal data, more shorter-and longer-term studies are necessary.

(d) Minipigs
Only one study described purely meniscus-related OA changes of the subchondral bone [4] (Table 5).In Yucatan minipigs, at 1 month after DMM, cartilage contact area decreased and concentrated at the cartilage-cartilage region [4], deep BV/TV decreased [4], and superficial Tb.Th increased [4].At 3 months, contact area, deep BV/ TV, and superficial Tb.Th all became normal [4].These changes might be due to an early transient loss of smaller trabeculae caused by increased loads, which is consistent with other early OA models [15].
Longer duration studies are needed to examine whether the minipig DMM model shows similar late OA subchondral bone sclerosis as humans.
At the anterior subregions after anterior partial medial meniscectomy, at 6 weeks, small osteophytes, increased subchondral bone plate porosity, unchanged subchondral bone plate thickness, decreased subchondral bone plate BMD; distinct loss and thinning of subchondral trabeculae were observed besides developing cartilage damage [43].In other subregions of the medial tibial plateau, such changes were minor and the subchondral bone plate unchanged [43].At 6 months, subchondral bone plate thickness increased, its porosity and BMD decreased, and large osteophytes occurred in the anterior subregions.The entire medial tibial plateau exhibited a strong loss of subchondral trabeculae (decreased BMD, and Tb.N, and increased Tb.Sp, and Tb.Th) [43,44].These data reveal a progressive loss of subchondral trabeculae, starting below the location of the meniscal injury at early and mid-term OA, reflected in disrupted correlations of microstructural osteochondral parameters.
Total meniscectomy, MMT and DMM all induced largely similar cartilage damage at 3 months [12].Differences included more anterior and focal cartilage lesions in DMM versus more widespread lesions.Osteophyte formation and subchondral bone plate thickness increased after total meniscectomy and MMT [12].In sum, early and mid-term ovine OA development was observed in the studied 1.5-9 months' time-frame.It is characterized by a local deterioration of the subchondral bone with osteophyte development, increased subchondral bone plate thickness, loss of bone volume and trabeculae, and decreased mineralization affecting primarily the compartment with compromised meniscal integrity, mostly independently of the applied technique (Table 6, Fig. 3g-k).Importantly, studies resembling human late OA by examining such ovine subchondral bone changes with longer follow-up time (several years) are noticeably lacking.

Discussion
The most important finding is the spatio-temporal pattern of subchondral bone remodeling: Changes in the subarticular spongiosa occur shortly before those of the subchondral bone plate.These early alterations involve a degradation of the trabecular elements, reduction of their number, loss of bone volume and reduced mineralization.Soon thereafter, the previously normal subchondral bone plate becomes thicker.Its porosity first increases, then decreases.Other essential conclusions are that: (1) Only early or mid-term time points were presented.(2) Larger joint sizes allow reporting higher topographical details. (3) The most frequently reported bone parameters were BV/TV (61%), BMD (41%), osteophytes (41%) and subchondral bone plate thickness (39%).( 4) Subchondral bone plate microstructure is not comprehensively characterized.(5) Microstructure of the subarticular spongiosa is well described.
Out of the 5367 identified meniscus-related OA studies, only 521 (9.7%) mentioned subchondral bone in its title or abstract, and out of them only 23 (0.4%) fulfilled the criteria to report subchondral bone characteristics in the surgical protocol solely based on meniscal damage in guinea pigs, rabbits, dogs, minipigs, and sheep.Data on dogs and minipigs were scarce with only a few published studies.For dogs, this might be due to the public perception being companion animals, the often complex ethical approval processes, and their difficult and costly management [41].Minipigs require specialized husbandry and food, and they are considerable less docile than sheep, and the miniature strains, more suitable for OA research than large agricultural pigs, are possibly less widely available in some countries [41].No data from horses or goats were identified.Some large animal studies reported their results in high topographic details, usually examining only 1 or 2 time points.In rabbits, contrastingly, 3-5 end points, covering a broader time scale from early to mid-term / late OA were sometimes reported.Only a low percentage (0.4%) of studies report subchondral bone characteristics.The number of such studies did not change recently for the examined larger species.Among them, rabbits are most frequently used (39%).OA is induced mostly in the medial compartment (87%), in a unilateral study design (61%), in the right (57%) knee, by total meniscectomy (48%).Micro-CT was only selected in 39% of the studies (histology: in 44%), although it is the most capable and recommended [8] method to analyze bone structure at high detail.
Eligible work presented only early or mid-term time points of OA development.Studies reporting more severe damage in the region below a meniscus lesion confirm the spatio-temporal pattern of subchondral bone remodeling induced by a meniscus injury [43,44].At the site of the injury, osteophytes appear relatively soon.Remarkably, changes in the subarticular spongiosa appear slightly before subchondral bone plate alterations, as ovine data suggests.These early alterations are characterized by a degradation of the trabecular elements and reduction of their number (decreased Tb.N), loss of bone volume (mostly decreased trabecular BV/TV, increased Tb.Pf and Tb.Sp), and reduced mineralization (BMD, TMD).Soon afterwards, the previously normal subchondral bone plate becomes thicker.Its porosity, a parameter negatively associated with sclerosis, first in-, then decreases (Fig. 4).
Methodological issues were also identified (Table 7).A detailed, quantitative 3D microstructural assessment of the subchondral bone is still not a general practice (performed in ~ 1/5 of the studies), limiting our knowledge about subchondral remodeling.Although cartilage analyses are relatively well standardized, similar methodical guidance of bone evaluation is absent.Such data would be needed to allow for clinically relevant and comparable conclusions in distinct model species and time-points.A standardization of the evaluation techniques and the reported parameters could be achieved for example by using the gold-standard micro-CT, reporting a minimum parameter set of BV/TV, Tb.Th, Tb.Sp, and Tb.N of the subchondral trabecular bone (as recommended in the classical paper of Bouxsein et al.) [8], together with subchondral bone plate thickness and osteophytes.These parameters can be reliably and accurately detected also in patients with clinical CT [40], allowing for a direct comparisons with the animal models.Still, only 22% of all examined studies report these 4 recommended trabecular parameters, and only 9% of all studies present the extended parameter set including subchondral bone plate thickness and osteophytes.
Definition of analysis volumes of interests (VOI) also needs standardization, by constantly separating the subchondral bone plate from the trabecular bone VOIs.Yet, many studies reported them together, even though separation is possible in all examined species [39].This appears especially important because simultaneous different direction of numerical changes in the two bone regions, observed commonly in various models (e.g. an increased BV/TV in In early human and late large animal OA related to traumatic meniscal injuries, structural subchondral data are largely absent that could provide crucial information about the temporal trajectory of changes.As many parameters decrease in early OA below normal, and increase above it in  late OA (or vice versa), a direct comparison of the results is not feasible without a clear definition of "early", "mid-term", and "late" OA stages within each species.Depending on OA stage, an increase, decrease or no difference of a certain parameter vs. normal is also possible and may reliably mirror the actual disease stage.Combining multiple microstructural parameters may identify a phenotypical fingerprint of each stage of the disease.Thus, detailed longitudinal studies with identical surgical protocols, multiple (early, mid, late) time points, appropriate normal/sham controls, and reliable and exhaustive structural analyses are required in all species.
Limitations include the absence of large animal microstructural data on late OA, complicating the comparability with humans.They are required in the future.High quality longitudinal studies revealing subchondral bone microstructural changes following meniscal injuries are unavailable, but needed to determine which animal model species represents best the human condition most faithfully.By using in vivo micro-CT or multiple termination time points, comparable longitudinal animal data could also be collected.While the meniscus tear is traumatic in nearly all animal models, the more common clinical situation of degenerative tears needs more attention [30].However, degenerative lesions of the meniscus will be difficult to imitate in animal models.
In sum, changes in the subarticular spongiosa have a short temporal priority over those of the subchondral bone plate.These early alterations involve a degradation of the trabecular elements, reduction of their number, loss of bone volume and reduced mineralization.Soon thereafter, the subchondral bone plate becomes sclerotic; its porosity first increases, then decreases.The specific human topographical pattern of a thinner subchondral bone plate in the region below both menisci is present solely in the larger species (partly in rabbits), but absent in rodents, an important fact to consider when designing animal studies examining subchondral consequences of meniscus damage.Large animal models are capable of providing high topographical detail, suggesting that they may represent suitable study systems reflecting the clinical complexities.Future studies need to assess the subchondral bone in a standardized fashion.Comparative longitudinal studies investigating its microstructure in early, mid-term, and late stages with appropriate normal controls in all larger animal species will allow addressing clinically relevant questions about fundamental determinants of subchondral bone remodeling in knee OA caused by meniscal injuries.

Fig.
Fig. 1 Flowchart of the systematic literature search resulting in n = 23 eligible papers, evaluated in the study

Fig. 2
Fig. 2 Summary of the n = 23 papers evaluating the subchondral bone in animal models of OA induced by meniscal injuries.a Histogram showing the number of eligible papers in 4-year intervals and the most important subchondral bone b evaluation methods and c parameters most frequently reported.Note that the cumulative percentages within the graphs may not be equal to 100% due to some studies reporting multiple techniques or parameters.d Reported time points

50 100Fig. 3
Fig. 3 Numbers and ratios of studies reporting the directions of changes of the individual bone microstructural parameters at different time points in multiple species.Stacked column diagrams showing OA-related changes of the bone microstructural parameters following a guinea pig medial meniscal transection (MMT) at 12 weeks and guinea pig destabilization of the medial meniscus (DMM) at b 1 and c 3 months, rabbit DMM at d 8, and at e 13-40 weeks, f dog DMM

Fig. 4
Fig. 4 Summary of the reported subchondral bone microstructural changes in early/mid-term OA. a 3-dimensional reconstructed micro-CT model of the subchondral bone plate and subarticular spongiosa showing the generally evaluated microstructural parameters.Representative safranin-O/fast-green stained histological sections of the medial tibial plateau of b a normal sheep and c a sheep 6 months after partial medial meniscectomy[43].Arrowheads point to characteristic subchondral bone microstructural alterations described in multiple animal models in early / mid-term OA, including(1)

Table 1
Bone microstructural parameters frequently used in micro-CT examinations

Table 3
Studies describing OA alterations of the rabbit subchondral bone following surgically induced meniscal injuries

Table 4
Studies describing OA alterations of the canine subchondral bone following surgically induced meniscal injuries At the study design "other treatment groups" mean drug or surgical treatments irrelevant for the present review BV/TV percent bone volume; DMM destabilization of the medial meniscus; Tb.Th trabecular thickness

Table 5
Study describing OA alterations of the porcine subchondral bone following surgically induced meniscal injury

Table 6
Studies describing OA alterations of the ovine subchondral bone following surgically induced meniscal injuries

Table 7
Recommendations and considerations for future studies examining OA development in the subchondral bone following meniscal injury 2D 2-dimensional; 3D 3-dimensional; BV/TV percent bone volume; micro-CT microcomputed-tomography; OA osteoarthritis; Tb.N trabecular number; Tb.Sp trabecular separation; Tb.Th trabecular thickness; VOI volume of interest Aspect to consider Future studies Explanation Controls Add normal/sham controls Appropriate controls were missing in several publications, and these studies had to be excluded from the present review Surgical protocol Clear and detailed description needed Description of surgical protocols was often insufficient to clearly identify the applied method of OA induction and relate it to the clinical problem Longitudinal studies Investigate early, mid, late time points To compare OA progression in different surgical models and animal species vs. human OA, longitudinal studies with multiple time points are needed Selection of volumes of interests Separate the subchondral bone plate from the subarticular spongiosa Bone microstructural alterations may be opposite direction in the two subchondral bone regions, thus separate VOI selection is crucial Gold-standard evaluation technique Perform micro-CT Micro-CT is capable of accurate 3D evaluation of the subchondral bone in high spatial resolution in contrast to 2D methods such as histology Minimum set of micro-CT parameters Evaluation of subarticular spongiosa: BV/TV, Tb.Th, Tb.Sp, and Tb.N Recommended [8] for a comparable characterization of the subchondral bone across different studies and models Extended set of micro-CT parameters Evaluation of subchondral bone plate: thickness and osteophytes (in addition to the minimum set) Additionally to the minimum set of subarticular spongiosa parameters recommended by Bouxsein et al. [8], these additional subchondral bone plate parameters give a more comprehensive view of the entire subchondral bone