**Histopathological Signatures of the Femoral Head in Patients with Osteonecrosis and Potential Applications in a Multi-Targeted Approach: A Pilot Study**

## **Giovanna Desando 1,\*,**†**, Livia Roseti 1,**†**, Isabella Bartolotti 1,\*, Dante Dallari 2, Cesare Stagni <sup>2</sup> and Brunella Grigolo <sup>1</sup>**


Received: 7 May 2020; Accepted: 4 June 2020; Published: 6 June 2020

**Abstract:** (1) Background: Osteonecrosis (ON) of the femoral head is a disabling disease for which limited treatment options exist. Identifying therapeutic targets of its evolution could provide crucial insights into multi-targeted approaches. The aim of this pilot study was to assess the histopathological features of patients with non-traumatic femoral head (NTFH) and post-traumatic femoral head (PTFH) ON to produce a fresh vision for clinical use. (2) Methods: We got biopsies from patients with different ON stages, according to the ARCO system. Samples from multi-organ donors were used as controls. Histological and immunohistochemical evaluations were performed on the osteochondral unit. (3) Results: The PTFH group displayed several fibrotic reactions, a small stem cell pool and a lower international cartilage repair society (ICRS)-I score than NTFH, which instead presented intact cartilage similar to the controls. Immunostaining for collagen I and autotaxin confirmed these features in the PTFH group, which displayed top levels of MMP-13 involved in cartilage loss and reduced CB-2 in the underlying bone. Both groups manifested a similar pattern of apoptotic and pain mediators. (4) Conclusions: The different histopathological features suggest a multi-disciplinary and multi-targeted approach for ON. Further studies are necessary to measure the effect size to gain clinical evidence.

**Keywords:** osteonecrosis; osteochondral unit; tissue remodelling and repair; multi-targeted approach

## **1. Introduction**

Osteonecrosis (ON) of the femoral head is a progressive and disabling disease, affecting active patients between the third and fifth decade of life with a high burden on the healthcare system [1–3]. The pathogenesis of ON involves the interplay of genetic, local and metabolic aspects with a different incidence rate among men and women (ratio male/female: 4:1) [4–7]. It is possible to identify two major types of ON aetiologies. The aetiology in ON patients with post-traumatic femoral head (PTFH) ON includes previous traumatic events; non-traumatic femoral head (NTFH) ON aetiology includes corticosteroid use, alcohol abuse, obesity, autoimmune diseases, and immunosuppressive therapies [7–9]. The pathology reflects a dynamic course leading to femoral head collapse because of subchondral bone fractures and inadequate bone repair [10–12]. Skeletal complications and pain in ON patients occurs because of abnormal osteoblastic and osteolytic activities [13,14]. In particular,

ON patients showed impaired action of the OPG/RANKL/RANK signalling pathway [9,14]. Wang X. and his group observed a similar behaviour whereby ON patients displayed increased levels of osteoprotegerin (OPG), receptor activator of nuclear factor-kB (RANK), and its ligand (RANKL) genes in ON necrotic areas [15]. Beyond such classical signalling pathways, scientists have even started testing the role of cannabinoids (CN) in bone repair, as bone-marrow-derived osteoclasts and osteoblasts [16], MC3T3 E1 osteoblast-like cells [17] and osteocytes express CN receptors [18]. Beyond the critical role of bone in the ON setting, scientists have pointed great attention towards the articular cartilage, because of its close anatomic continuity and communication with the underlying subchondral bone [19,20]. In particular, several authors stressed the clinical significance of considering structural cartilage changes following the mechanical stress induced by the collapse of the subchondral bone in ON [21–23]. However, these changes contribute to altering the biomechanics of the joint leading to OA progression and finally to joint destruction.

There is no gold-standard treatment for ON because of the controversial results [24,25]. Selecting treatment options for ON depends on many factors, including the pathology stage, patients' age and health conditions, and lesion size and location [26,27]. In particular, treatments for ON management in the pre-collapse stage include non-surgical (weight control), pharmacological or biophysical techniques, and mesenchymal stromal cell (MSC)-based therapy [27–31]. Total hip arthroplasty is among the best therapeutic alternatives after the collapse of the femoral head. However, it has its disadvantages (e.g., infection, revision, and dislocation) [24,25]. Therefore, many scientists focused on femoral head regeneration. In 2002, Hernigou P. and his group reported promising results following MSC therapy in ON patients in the pre-collapse stage at 5 to 10 years of clinical follow-up [29]. A recent review summarised that MSCs could regenerate the necrotic area of the femoral head by injecting the suspension into the lateral artery of the circumflex or loading on carriers via core decompression and implantation [32]. Testing the crosstalk between cartilage and the underlying subchondral bone is essential to achieve global comprehension of physiological responses in ON disease. Differences between ON patients with various aetiologies could give first perspectives for tailored-based therapies. In this light, we conducted a small-scale preliminary (pilot) study aimed at evaluating histopathological features in two ON patient groups with NTFH and PTFH. Study design foresaw investigating several mediators modulating the osteochondral unit to get more knowledge of potential therapeutic targets for ON. In particular, we selected collagen I and autotaxin (ATX) to test the fibrotic reactions commonly causing poor mechanical properties and the limited capacity of MSCs to differentiate towards collagen type II in cartilage [33,34]. The axis ATX/lysophosphatidic acid (LPA) regulates collagen type I biosynthesis and plays essential functions in bone metabolism, thus resulting in an attractive molecular target [34,35]. We tested tissue destruction by selecting matrix metalloproteinases (MMP-13) and aggrecanases (ADAMTS-5), causing the proteolytic cleavage of collagens and the aggrecan protein [36]. As for apoptosis reactions, we tested active caspase 3, which is a well-known biochemical marker of both early- and late-stage apoptosis [37]. We tested the endocannabinoid receptor-2 (CB-2), as it is present in distinct cell types of the joint, like chondrocytes, bone cells, progenitor cells during osteoarthritis and rheumatoid arthritis with potential therapeutic implications [38]. We chose the neurotrophin nerve growth factor (NGF) and the nociceptive peptide substance P (SP) to test the pain response [39,40].

## **2. Materials and Methods**

#### *2.1. Patient Data and Surgical Procedure*

This study obtained approval from the Ethics Committee of IRCCS Istituto Ortopedico Rizzoli (Prot. gen. n. 26146 del. 31.10.2006). Eleven male patients with clinical and radiological signs of ON of the femoral head gave their informed written consent to this study. We selected patients according to specific inclusion and exclusion criteria. Inclusion criteria foresaw the enrolment of male patients with a mean age between 18–50 years and clear signs of hip osteonecrosis by MRI. Exclusion criteria foresaw the exclusion of patients showing metabolic diseases, rheumatoid arthritis, autoimmune and neurological disorders. In this study, five ON patients showed NTFH (mean age 37 ± 4; range: 30–43), whereas six ON patients had PTFH (mean age/SD: 30 ± 3; range: 25–37). The aetiology of the NTFH group included glucocorticoid treatment. ON patients included in the NTFH group did not report aetiology for alcohol abuse and autoimmune diseases; only one patient had a chronic bowel disorder. NTFH and PTFH patients underwent preoperative MRI and X-ray using the Association Research Circulation Osseous (ARCO) evaluation system [41]. This system considers the size of the necrotic lesion, its femoral head extent, and the joint involvement (Stage I: normal on X-ray and CT; Stage IV: the destruction of joint with secondary arthritic changes; A, B and C describe the extent of ON involvement: A: non-articular, B: medial; C: central). We reported a scientific diagram of the weight-bearing area of the femoral epiphysis where we harvested biopsies (see Supplementary Figure S1). Patients' femoral heads underwent surgical repair with a synthetic resorbable osteochondral scaffold plug. During the procedure, surgeons collected osteochondral biopsies of the lesions from the weight-bearing area of the femoral head with a 10-mm diameter through a 12–14-mm deep trocar. All samples were processed for histological and immunohistochemical analyses. We used the femoral head from three male multi-organ donors (mean age/SD: 35 ± 5; range: 30–40) as healthy controls. Multi-organ donors did not suffer from any musculoskeletal disease. We selected the donors through the bone bank program for tissue donation after the family' s donor consent. Femur harvesting was performed within six hours from asystole, and involved its excision and placement in Dulbecco-modified Eagle medium with L-glutamine, NaHCO3, and antibiotics, and storage at 4 ◦C.

#### *2.2. Histological Assessment*

Osteochondral samples were fixed with 10% buffered formalin and decalcified in 4% hydrochloric acid and 5% formic acid [42]. After processing with a graded alcohol series, specimens were embedded in paraffin. We tested proteoglycan and collagen content by staining tissue sections with 0.1% Safranin-O/0.02% Fast Green (Sigma Aldrich, St Louis, MO, USA). International Cartilage Repair Society (ICRS)-I score was used for evaluating the histological features [43]. This score considers six parameters: surface, matrix organization, cell distribution, cell viability, subchondral bone and cartilage mineralization. It has a range from 0 (presence of fibrous tissue) up to 18 (presence of healthy osteochondral tissue). We assessed necrotic bone lesions with the Ficat and Arlet classification system [44]. This system considers four types of bone necrosis. Its score ranges from 0 (slight disease) to 4 (severe disease). Six microscopic fields, spaced 20 sections, were assessed for each sample by two blinded investigators (GD, IB) with an Eclipse 90i microscope (Nikon, Melville, NY, USA).

#### *2.3. Immunohistochemical Analyses*

Analyses for collagen type I, caspase-3, MMP-13, ADAMTS-5, autotaxin, NGF, SP, and CB-2 were performed. After antigen retrieval with 0.1% proteinase (Sigma) at 37 ◦C for 20 min, the sections were blocked with 2% bovine serum albumin (Sigma) in phosphate-buffered saline for 30 min. Then, an incubation with human collagen type I (2 μg/mL; Chemicon International, Temecula, CA, USA), caspase-3 (5 μg/mL; R&D Systems), ADAMTS-5 (1 μg/mL, Abcam), MMP-13 (5 μg/mL, R&D Systems, Minneapolis, MN, USA), autotaxin (2 μg/mL; R&D Systems), NGF (1 μg/mL, Chemicon), and CB-2 (5 μg/mL; Novus Biologicals) was carried out. Specific negative controls were performed by omitting the primary antibodies or using an isotype-matched control while we stained nuclei with CAT hematoxylin (Biocare Medical). Six microscopic fields (100× magnification) were assessed for each sample by a blinded investigator with a semi-quantitative method. We firstly segmented cartilage and the subchondral bone for each marker by selecting zones apart from the tidemark, especially for the PTFH group where the tidemark was fragmented. Image acquisition and processing with an Eclipse 90i microscope (Nikon) and NIS-Elements Software were used for the image analysis of stained sections with the Hue/Saturation/Intensity (HSI) system. Hue (H) was assessed by setting the threshold for

positive pixels at 220 to 255. Ranges from 0 to 150 were threshold values for S and I. The measurement of positive cells and area for each marker was done on the entire osteochondral sample (10× objective lens) and expressed as a percentage of positive cells and area on a scale from 0 (no protein expression) to 100 (the highest protein expression).

## *2.4. Statistical Analysis*

Graph Pad Prism software was adopted for statistical analysis. The Kolmogorov–Smirnov test was used to test the data distribution. We used the Mann–Whitney U test for unpaired data to assess differences in NTFH and PTFH groups. We reported data in a scatter plot graph with mean ± standard deviation (SD). *p* < 0.05 was considered significant.

#### **3. Results**

#### *3.1. Radiographic Assessment*

According to the ARCO osteonecrosis classification system, patients with NTFH and PTFH displayed different stages of ON. The NTFH group reported three patients with III C stage and two patients with IV C. Patients with III C stages showed clear signs of ON and separation of the subchondral bone from the necrotic cancellous bone. Patients with IV C revealed joint space narrowing following the femoral head collapse. The PTFH group included four patients with IV C and two with IV B, which reported bone fracture and subsequent arthritic changes.

#### *3.2. NTFH and PTFH Groups Displayed Di*ff*erent Histological Features in Cartilage and Bone*

The control group showed a regular cartilage surface, adequate cell distribution and rich proteoglycan content. Bone tissue was also well-structured with osteocytes embedded in the bone matrix and trabecular spaces containing bone marrow and blood vessels (Figure 1a). The NTFH group showed a regular cartilage surface with small discontinuities and good proteoglycan content. The extracellular matrix of specimens with radiographic IV C stage displayed a reduced number of cells, some cell clones, and tidemark discontinuities. Bone tissue showed a low number of osteocytes and necrosis of the bone marrow in the trabecular spaces. No inflammatory reactions were, however, present (Figure 1a). The PTFH group exhibited a typical fibrocartilaginous aspect with several cracks in the superficial zone, and an altered cell arrangement with round cells interposed in the extracellular matrix. The tidemark displayed non-continuous areas with cells migrating from the subchondral bone towards the cartilage. The PTFH group showed trabecular spaces containing fibroblasts, blood vessels, and osteoclasts but no inflammatory infiltrate (Figure 1a). The NTFH group showed a higher ICRS-I score than the PTFH group (*p* < 0.05), reporting mean values of 13.1 ± 0.6 and 7.8 ± 0.9, respectively (Figure 1b). The PTFH group exhibited a worse histological aspect, different from the control group, which reported a mean value of 16.8 ± 0.6 (*p* < 0.001) (Figure 1b). As for the cartilage parameters, the extracellular matrix and the tidemark showed better organization in NTFH rather than in the PTFH group (*p* < 0.001) (data not shown). The Ficat classification system gave evidence of more degenerative changes in the subchondral bone marrow of PTFH group when compared with NTFH group but with no statistical evidence (Figure 1b).

**Figure 1.** (**a**) Safranin-O/Fast Green staining of cartilage and subchondral bone from healthy controls, and two groups of patients with non-traumatic femoral head (NTFH) ON (*n* = 5), and post-traumatic femoral head (PTFH) ON (*n* = 6). Red: proteoglycans; green: collagen. Scale bar: 50 μm. IG isogenic groups: CL cell clusters; L: empty lacunae; FIS fibrillation processes; HC hypertrophic chondrocytes; T tears; Bo bone tissue; BM bone marrow; AT adipose tissue; BV blood vessels; FIB fibrous tissue; OC osteoclasts; E erythrocytes. (**b**) Graphical representation of ICRS and Ficat scores of the osteochondral unit of control, NTFH and PTFH groups. Data are reported in a scatter plot graph with mean ± standard deviation (SD). \* *p* < 0.05: NTFH versus PTFH; \*\*\* *p* < 0.001: PTFH versus the control group.

#### *3.3. PTFH Group Displayed a Higher Expression of Fibrotic Markers than NTFH Specimens*

The control group showed low protein expression for type I collagen in cartilage, whereas we noticed a high percentage of this marker in the underlying subchondral bone (Figure 2a,b). Fibrosis reactions in terms of the presence of type I collagen were more robust in the cartilage of the PTFH group compared to the NTFH group (Figure 2a,b). This latter group showed higher type I collagen expression compared to the control (*p* < 0.05) and NTFH groups (*p* < 0.05). As for autotaxin, the two ON groups did not show any difference; however, the PTFH group reported a higher amount of this marker compared to the control group (*p* < 0.05). The NTFH group showed mild protein expression for both collagen type I and autotaxin, especially in the articular cartilage.

**Figure 2.** (**a**) Immunostaining for collagen type I and autotaxin of control, NTFH and PTFH groups. Scale bar: 50 μm. Black arrows: positive areas for markers. (**b**) Graphical representation of quantitative measurements for collagen type I and autotaxin in control, NTFH and PTFH groups. Data are reported in a scatter plot graph with mean ± SD. Collagen type I: \* *p* < 0.05: Control versus PTFH group; \* *p* < 0.05 NTFH versus PTNH group. Autotaxin: \* *p* < 0.05: Control versus PTFH group.

We did not find any difference for caspase-3 between the NTFH and PTFH groups. Both specimens displayed higher cell positivity for this marker when compared to the control group (Figure 3a,b). Regarding CB-2, a molecule involved in bone remodelling and pain responses, the control group displayed moderate expression, especially near bone marrow spaces, and at a lesser extent in the osteocytes. The NTFH group revealed higher protein expression for CB-2 in chondrocytes than in osteocytes and bone marrow precursors within the subchondral bone. The PTFH group displayed a similar behaviour (Figure 3a,b).

To assess the breakdown of the extracellular matrix, we analysed ADAMTS-5 and MMP-13 as catabolic markers. All cartilage and bone specimens showed low levels of ADAMTS-5, with no difference between the two ON groups (Figure 4a,b). As for MMP-13, the PTFH group displayed a higher cell positivity in the middle and deep layers of articular cartilage when compared to the control and NTFH groups (*p* < 0.05) (Figure 4a,b).

**Figure 3.** (**a**) Immunostaining for caspase-3 and CB-2 of the control, NTFH and PTFH groups. Scale bar = 50 μm. Black arrows show positive cells. (**b**) Graphical representations of the percentage of positivity for caspase 3 and CB-2. Data are expressed in a scatter plot graph with mean ± standard deviation (SD). Caspase 3: \* *p* < 0.05: Control versus NTFH group; \* *p* < 0.05: NTFH versus PTFH group.

**Figure 4.** *Cont.*

**Figure 4.** (**a**) Immunostaining for ADAMTS-5 and MMP-13 in osteochondral tissues from the control, NTFH and PTFH groups. Scale bar = 50 μm. Black arrows show positive cells. (**b**) Graphical representations of the percentage of positivity for ADAMTS-5 and MMP-13. Data are expressed in a scatter plot graph with mean ± standard deviation (SD). MMP-13: \*\* *p* < 0.001: Control versus PTFH group; \* *p* < 0.05: NTFH versus PTFH group.

We analysed NGF and the substance P (SP) nociceptive fibre pattern to test the pain response. Both NTFH and PTFH groups showed higher immunostaining for NGF in the cartilage rather than in the underlying subchondral bone (*p* < 0.05) (Figure 5a,b). Both ON groups displayed an increased SP positivity near cell clones of the articular cartilage and hypercellular and fibrotic areas of the subchondral bone (Figure 5a,b).

**Figure 5.** (**a**) Immunostaining for NGF and SP in osteochondral tissues from control, NTFH and PTFH groups. Scale bar: 50 μm. Black arrows: positive areas for the selected markers. (**b**) Graphical representation of quantitative measurements for NGF and SP in control, NTFH and PTFH groups. Data are expressed in a scatter plot graph with mean ± standard deviation (SD).

#### **4. Discussion**

The search for ON therapies capable of controlling the joint microenvironment is a unique challenge [1,2]. A holistic view of biological responses during ON can produce insights for generating efficient strategies [22]. Our pilot study showed different histopathological features in two groups of ON patients by opening up interesting biological perspectives. All specimens with NTFH and PTFH forms displayed basal integration with the underlying bone. However, PTFH samples showed noticeable fibrotic reactions and a reduced stem cell pool in the bone marrow. Interrupting blood supply likely contributed to mediating the death of bone and marrow cells. As the bone supports the exchange of nutrients with cartilage by facilitating joint force distribution, bone alterations might speed up cartilage degeneration [22,45]. The subchondral bone changes may lead to misalignment of the articulating surfaces with consequences on the mechanosensory cells in the bone [45]. Thus, first-line approaches for shifting the bone matrix turn-over are essential to avoid progressive degenerative changes in cartilage tissue. Bisphosphonate treatment could restore the balance between bone resorption and formation [26]. Biopsies from patients with post-traumatic aetiology reported several OA features, including impaired extracellular matrix and vascular infiltration in cartilage because of the mechanical stresses. We determined the relevance of matrix composition and tidemark presence using a semi-quantitative analysis with the ICRS score, by observing different histological scores in the PTFH and NTFH groups. This latter displayed similar results to the control group by reporting intact cartilage but several changes in the subchondral bone. These findings are in line with other studies, which showed intact articular cartilage in the NTFH group with corticosteroid treatments [22]. The two groups even exhibited distinct protein expression for mediators modulating fibrotic and catabolic responses.

Protein assessment for collagen type I and autotaxin corroborated the fibrotic aspect in both cartilage and bone from the PTFH group. The autotaxin–lysophosphatidic acid (LPA) axis is emerging as a critical regulator in various biological responses [46]. Autotaxin exerts a fibrotic activity by catalysing LPA, which promotes fibrosis responses by regulating collagen type I biosynthesis [34,47]. Moreover, autotaxin controls the fusion and bone resorption capacity of osteoclasts [35]. Another peculiar feature observed in the cartilage and bone tissues from the two ON aetiologies was the protein expression for MMP-13. It is a molecule which stimulates collagen and proteoglycan degradation in cartilage and bone [48]. Along this path, Grassel S. et al. provided first evidence of MMPs in the ON setting. ON patients with PTFH showed increased gene expression levels of MMP-2 and a low amount of the tissue inhibitor of metalloproteinases (TIMPs) [49]. The increased protein levels of MMP-13 in the PTFH group may likely depend on the up-regulated activity of chondrocytes in producing MMPs after stress forces. These findings open preliminary biological insights into considering MMPs and fibrotic markers as attractive therapeutic targets by exploiting future technological advances like CRISPR-Cas9 [50]. Combining current ON strategies with small molecule inhibitors to block collagen type I, MMP-13 and ATX could improve the success of PTFH treatment [47,51]. Regardless of the ON aetiology, samples from both groups displayed a moderate positivity for caspase-3, showing higher values than the control. Beyond the classical signalling pathway, we considered the role of CB-2 involved in the endocannabinoid system. Jiang S. et al. reported that this system promotes specific signalling pathways in response to pathogenic events to launch repair processes [52]. In specimens from both groups, we showed positive immunostaining for CB-2, especially in cartilage, with higher values than healthy controls. We noticed a low expression of CB-2 in the bone marrow niche, likely because of the small number of osteoprogenitors in ON patients.

Like in ON disease, femoral heads from patients with the atrophic form of osteoarthritis (OA) display bone marrow lesions with numeric, topographic and functional variations of MSCs [53,54]. Bone marrow perturbations reduce the bone repair and remodelling activities of MSCs and lead to damage of the overlying articular cartilage [55,56]. Targeting both cartilage and bone turn-over and the crosstalk between their cell types may be a valuable approach [56]. In this light, there is abundant evidence of the powerful effects of MSCs for promoting osteochondral repair and inhibiting

inflammatory and fibrotic reactions in the OA environment [57–59]. Optimal therapeutic approaches might envisage several interventions, at varying stages of OA and ON disorders and selecting specific patient features. Restoring the stem cell pool, especially in the PTFH group, through cell-based therapy, could be a feasible intervention. The remarkable bone–cartilage interface imbalance of this patients group, simultaneously with the low stem cell pool, may benefit from the differentiation and paracrine properties of MSCs. Hernigou P et al. reported promising clinical results using autologous bone marrow transplantation in ON cases by highlighting the biological and therapeutic value of this strategy [30]. Along the way, several clinicians have observed the combination of autologous bone marrow concentrate with core decompression contribute in a significant manner to decelerating ON progression by limiting total hip arthroplasty [60,61].

Finally, this study considered pain implications in the biopsy samples from two ON patients focusing on NGF and substance P; this latter is involved in the onset of inflammatory processes and pain transmission [52]. Specimens from both groups manifested pronounced immunostaining for NGF in cartilage. Beyond its role as a neurotrophic factor, NGF impairs the migratory and matrix remodelling activities of cartilage and stem progenitor cells [62,63]. More in-depth investigations are crucial to assess the effect size and produce clinical evidence, as the small number of ON cases and heterogeneity are the major limitations of the present study. However, this pilot study gave indications to enhance the benefit of current ON approaches by merging themselves with fibrotic inhibitors and choosing a multi-disciplinary and targeted strategy to both cartilage and bone in the PTFH group.

### **5. Conclusions**

Sustainable multi-disciplinary strategies can represent valid tools to tackle complex pathologies, with tremendous impact in the clinical decision-making framework. In our pilot study, the different histopathological features of NTFH and PTFH groups would suggest a multi-disciplinary and multi-targeted approach for both cartilage and bone tissues. Hypothetically, restoring the stem cell pool in the subchondral bone from the PTFH group could be fruitful in supporting tissue regeneration. Successful clinical results could be obtained by hindering the fibrotic and catabolic responses at the level of cartilage and bone in the PTFH. Further studies are necessary to measure the effect of size to gain clinical evidence.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-3417/10/11/3945/s1, Figure S1: Scientific diagram.

**Author Contributions:** Conceptualisation, L.R. and G.D.; methodology, G.D.; C.S. and I.B.; software, G.D.; formal analysis, G.D. and D.D.; investigation, L.R., D.D. and C.S.; data curation, G.D., C.S. and I.B.; writing-original draft preparation, L.R. and G.D.; writing-review and editing, G.D., L.R., I.B., C.S., D.D., and B.G.; supervision, B.G.; project administration, B.G.; funding acquisition, B.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Italian Ministry of Health, 5× 1000 Funds anno 2016. "Malattie osteoarticolari: fisiopatologia e strategie terapeutiche innovative".

**Acknowledgments:** The authors wish to thank Patrizia Rappini and Martina Rocchi for their technical and scientific help.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

## **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Review* **New Insights on Mechanical Stimulation of Mesenchymal Stem Cells for Cartilage Regeneration**

## **Silvia Ravalli 1, Marta Anna Szychlinska 1, Giovanni Lauretta <sup>1</sup> and Giuseppe Musumeci 1,2,3,\***


Received: 6 April 2020; Accepted: 22 April 2020; Published: 23 April 2020

**Abstract:** Successful tissue regeneration therapies require further understanding of the environment in which the cells are destined to be set. The aim is to structure approaches that aspire to a holistic view of biological systems and to scientific reliability. Mesenchymal stem cells represent a valuable resource for cartilage tissue engineering, due to their chondrogenic differentiation capacity. Promoting chondrogenesis, not only by growth factors but also by exogenous enhancers such as biomechanics, represents a technical enhancement. Tribological evaluation of the articular joint has demonstrated how mechanical stimuli play a pivotal role in cartilage repair and participate in the homeostasis of this tissue. Loading stresses, physiologically experienced by chondrocytes, can upregulate the production of proteins like glycosaminoglycan or collagen, fundamental for articular wellness, as well as promote and preserve cell viability. Therefore, there is a rising interest in the development of bioreactor devices that impose compression, shear stress, and hydrostatic pressure on stem cells. This strategy aims to mimic chondrogenesis and overcome complications like hypertrophic phenotyping and inappropriate mechanical features. This review will analyze the dynamics inside the joint, the natural stimuli experienced by the chondrocytes, and how the biomechanical stimuli can be applied to a stem cell culture in order to induce chondrogenesis.

**Keywords:** mesenchymal stem cells; tissue engineering; chondrogenesis; osteoarthritis; bioreactor; mechanical stimuli; physical stimulation; compression; shear stress; hydrostatic pressure

## **1. Introduction**

Scientific minds have always expressed a stubborn yearn to master the creation of artificial biological elements in order to emulate the delicate machinery of the human body. The idea of being able to control cell destiny and replace malfunctioning parts of the body with brand new tissues and organs has led to remarkable progress in biological fields such as tissue engineering and regenerative medicine. A promising application of this concept relies on the use of stem cells to originate functional and specialized tissues [1]. Mesenchymal stem cells (MSCs) constitute a specific subtype of multipotent stem cells, which can differentiate into a variety of cell types and offer the advantage of obtaining pure stem cell populations [2]. One of the most challenging ambitions in regenerative biomedicine is to restore damaged articular cartilage, as it is one of the most challenging tissue types to heal by virtue of its anatomical and structural complexity. More precisely, the avascular, alymphatic, and aneural nature of the cartilage, combined with the fact that it is characterized by the chondrocyte cell type only, limits its ability to self-repair [3].

Conventional treatment and modern therapies to treat this singular tissue, ranging from injections to surgical procedures, still suffer, in many instances, from wide variation in clinical outcome, complications, specificity, and effectiveness and a lack of well-grounded long-term reliability [4]. Joint injection is an easy and minimally invasive procedure for the delivery of MSCs as anti-inflammatory mediators and immune-modulating factors [5]. Pain relief, improvement in articular function, and regeneration of cartilage following this treatment have been observed in a growing number of studies [6–8]. However, the efficacy of this procedure is controversial, since after injection, cells might not survive or remain in situ in the long term [9,10]. Satué et al., observed MSCs migrating and engrafting into the damaged cartilage as early as the first day after injection [11]; in contrast, in another recent study, MSCs were mostly found in the synovium but not in the cartilage surrounding the defect [12]. Disappearance of the injected cells within the joint may be due to failure in the extracellular matrix (ECM) attachment mechanism [12]. Although intra-articular injection of MSCs appears to be safe during the short term [13], further investigations, such as randomized controlled trials, are necessary to explore long-term adverse events and reduce the heterogeneous nature of the studies in terms of design, cell number, exogenous factors, and administration protocols.

Tissue engineering has raised interest as a reasonable approach to manage pathologies like osteoarthritis [1]. The latter is a degenerative disease characterized by progressive loss of articular cartilage, synovial inflammation, osteophyte formation, and joint space narrowing that lead to overall stiffness, pain, and loss of mobility of the affected joint. The bases of this escalating damage rely on a compromised balance between anabolic and catabolic mechanisms, which can be consequent to several risk factors like ageing, muscle atrophy, metabolic disorders, inflammatory conditions, injuries and overload, or wrong biomechanics of the joint [14]. This review aims to explore cartilage dynamics and understand what kind of biophysical principles will be beneficial for the engineering-based treatment of degenerative and rheumatic diseases of the joints.

## **2. Chondrogenesis In Vitro**

The essence of regenerative therapy lies in the use of stem cells, bioactive molecules, biomaterials, and their combinations [15]. Physiological chondrogenesis is regulated by specific cytokines and transcription factors, such as the transforming growth factor beta (TGF-β), fibroblast growth factor (FGF), and insulin-like growth factor 1 (IGF-1) [16]. As another example, SRY-related high-mobility group box gene 9 (SOX-9) is a transcription factor essential for chondrogenesis processes [17] and also to avoid dysregulated chondrocyte hypertrophy [18]. Commonly, MSCs undergo chondrogenic differentiation, recognizable by increasing in specific ECM components, when in the presence of TGF-β. This pathway consists of the proper formation of the receptor complex that requires type I and type II serine/threonine kinase receptors and intracellular effectors. The type I receptor family counts seven members which phosphorylate different small mother against decapentaplegic (SMAD) proteins. Transcriptional activation of chondrogenic genes, like those encoding type II collagen and aggrecan, is mediated by the involvement of the type I receptor 5 (ALK 5) and subsequent phosphorylation of SMAD 2/3. Otherwise, TGF-β could bind ALK 1 and activate the SMAD 1/5/8 pathways, resulting in hypertrophy-related gene expression [19]. Nevertheless, one of the clinical limits of this kind of chondrogenic differentiation, based on the TGF-β pathway, relies on the risk of incurring undesired and premature hypertrophy entrance of MSCs [20]. Considering this scenario, a more sophisticated and authentic approach to differentiation is offered by understanding the mechanical forces naturally experienced by chondrogenic progenitors and applying them, even in the absence of exogenous growth factors, to stem cells in order to mimic the same environment within the joint.

#### **3. Influencing the Mechanical Environment**

The characteristic structure of the articular cartilage is the result of the dynamic processes that occur within the joint. Chondrocytes experience biomechanical stimuli like compression, shear stress, and hydrostatic pressure [19]. These forces are perceived as a shifting of currents, electrical fields, or changes in osmolarity and so converted into intracellular signals, influencing mechanisms like transcription, exocytosis, and activity of Na+/K+-ATPase [21]. Growth factors such as TGF-β, IGF, and bone morphogenetic proteins (BMP)-2,-4,-7 are necessary to stimulate chondrogenic processes and require the presence of calcium ions (Ca2+) to regulate cell functions, such as the synthesis of extracellular matrix components. Physical stimuli have been associated with the regulation of Ca2<sup>+</sup> entry, primarily through voltage-operated calcium channels (VOCCs), transient receptor potential (TRP) channels, and purinergic receptors [22]. Furthermore, VOCC inhibitors have shown to reduce cartilage degradation and the progression of osteoarthritis [23], suggesting their importance in both physiological and pathological milieux. TRP channels, such as TRP vanilloid 4 (TRPV4), which have been linked to upregulation of the SOX9 pathway, or TRPC1, able to guide chondrogenesis in stem cells and regulate the activity of other voltage-dependent ion channels, are also highly involved. Inhibitors for these two receptors, 2-aminoethoxydiphenylborane (2-APB) and Ruthenium Red, have been shown to prevent MSC chondrogenesis induced by pulsed electromagnetic fields (PEMFs) [24]. The structure of the articular cartilage comprises several layers: First, a thin superficial zone, where the collagen fibers are aligned parallel to the surface and the chondrocytes are numerous and flattened. This layer is in tight contact with the synovial fluid, and it can resist shear stresses. Under the first zone, there is a transitional zone which contains especially proteoglycans (PGs) employed for compressive resistance. Here, the collagen is arranged obliquely. Just below, the deep zone can cushion the compressions and presents collagen fibers organized perpendicular to the surface, high concentrations of proteoglycans, and columns of cells. Lastly, the tide mark separates this last zone from the calcified cartilage [25] (Figure 1).

**Figure 1.** Structure of the articular cartilage. The structure of the articular cartilage comprises several layers: First, a superficial zone, where the collagen fibers are aligned horizontally, the chondrocytes are numerous and flattened, and the proteoglycan content is low. Under the first zone, there is a transitional zone which contains high levels of proteoglycans, randomly arranged collagen, and round-shaped cells. Just below, the deep zone presents vertically aligned collagen fibers, a high concentration of proteoglycans, and columns of cells.

For the sake of simplicity, this tissue could be imagined like a biphasic model: one phase is represented by the interstitial water that permeates the other phase, made up of the solid components of the ECM (collagen, PGs and other proteins). These two phases are extremely interdependent in terms of functionality and biomechanics. Compression is one of the forces experienced by cartilage, which leads to an internal increase of hydrostatic pressure of the aqueous phase. As a result, the water leaks from the ECM towards the capsule, but the structure of the cartilage will remain unaltered thanks to the PG component. More precisely, the effect of uniaxial compression will be compensated by the tensile stiffness generated by the repulsive forces between the negatively charged carboxylic or sulfonic groups of the glycosaminoglycans (GAGs). When the pressure ceases, the water is again attracted inside the interstices. Tissue compressibility under load reaches even a millimeter, but, when the spring back is not able to compensate hard and long compressions, the structure can be damaged [26]. Shear or rotational stress, which could be defined by the change in thickness with respect to the original height, is another force generated by joint movement and is caused by the tangential friction of synovial fluid on the surface. This movement allows synovial fluid to nourish the cartilage, transport waste materials, and keep the chondrocytes metabolically active with a mechanism of diffusion and fluid convection [27]. Under this condition, the collagen network is the viscoelastic component of the tissue that exhibits cushioning ability. The collagen concentration is directly proportional to resistance to shifting [28]. Weight-bearing articular cartilage of the hip and knee daily experience stress amplitudes from 0.5 to 7.7 MPa and average compression of 13% [29–31]. Chondrocytes show selective responses to various mechanical stimuli. Indeed, dynamic stresses are able to improve the production of ECM components, while static compressions do not lead to great achievements in tissue engineering constructs. Mechanical stimulation triggers those pathways that culminate in maintaining functional ECM in order to provide substantial physical stability against the stresses to which the cartilage is subjected. It is a feedback cycle in which the mechanical stress influences the production of those components which sustain the stress itself. It is not surprising that, in fact, an unbalanced step in this cycle could pave the way to a pathological mechanism which could, in the end, lead to the onset of osteoarthritic features. If physiological stimulation fails, following, for example, sedentary habits, or it exceeds the ability of the tissue to sustain it, e.g., excessive mechanical loading [32], the chondrocytes will miss most of the input to produce the new ECM, resulting in unbalanced homeostasis. It is highly recommended to patients who suffer from early osteoarthritis, who are able to conduct physical exercise, to encourage the movement of the diseased articulation in order to stimulate the restoration of the physiological cycle, which may lead to improvements in biochemical disorders. This dynamic environment should be considered in tissue engineering approaches in order to realize as realistic a construct as possible. The challenge proposed is to move from a purely biological view of the natural cell to an engineered one.

Stem cells are studied as structures able to sense and transmit physical stimuli, translating them into biological and mechanical responses, since they have greater mechanical sensitivity than adult cells [33,34]. As already mentioned, ion channels are paramount in triggering those signaling pathways which lead to matrix turnover and homeostasis, and an intracellular increase of Ca2<sup>+</sup> levels has been considered as one of the stem cell responses to mechanical load. Sequestration of calcium ions and inhibition of VOCCs and other channels have been shown to attenuate the effects of mechanical stimulation. More specifically, during physical stimulation of MSCs, Ca2<sup>+</sup> is known to be involved in the activation of pivotal transcription factors leading to chondrogenic differentiation [20]. In the next section, some interesting works on the topic are presented. In some of those, no exogenous growth factors were added during the experiment. Hence, chondrogenic differentiation was achieved exclusively as a result of load applications. Figure 2 illustrates the different types of mechanical forces applied in these experiments.

**Figure 2.** Types of mechanical stimuli experienced by the cells. Stem cell cultures, used in tissue engineering, undergo biophysical stimulation in order to promote chondrogenesis and recreate the dynamic environment within the joint.

#### **4. Practical Applications**

To reveal the physiological effect of mechanical stimulation on stem cells, loading machines were designed to be able to apply different types of mechanical forces with different intensity, duration, and frequency.

Cochis et al. [35] cultivated MSCs in a methylcellulose solution retained within a porous polyurethane matrix to evaluate the suitability of the matrix in supporting mechanically induced chondrogenesis of the cells in the absence of exogenous factors. The composite underwent a combination of compression and shear forces by the use of a bioreactor which applied compressive and rolling movements through a ceramic ball. This ball compressed the scaffold dynamically at 1 Hz, resulting in a strain amplitude of 10–20% of its height, and provided shear stress with oscillations perpendicular to the scaffold axis of ±25◦ at 1 Hz. After 21 days, the samples were analyzed by reverse transcription polymerase chain reaction (RT-PCR), biochemical, histochemical, and immunofluorescent assays, and the authors concluded that the physical stimulation led to the activation of TGF-β and to accumulation in the surrounding matrix of glycosaminoglycans and type II collagen, confirming the chondrogenesis process.

Another study on the involvement of mechanical forces in the destiny of adipose-derived mesenchymal stem cells (ADSCs) by Zhang et al. [16] highlights the role of dynamic compression in combination with exogenous SOX-9 on chondrogenesis. ADSCs were seeded in 3D porous polylactic-co-glycolic acid (PLGA) scaffolds. Gradual, meaning with a unique structure at each level, and uniform scaffolds were subjected, with the use of a bioreactor, to sequential uniaxial compressions of 5–10% strain amplitude and frequency of 0.1 Hz. The authors analyzed the morphology of the ADSCs and the ECM deposition within the scaffolds by scanning electron microscopy (SEM), noticing that ECM accumulation by ADSCs, on gradual scaffolds and in the presence of SOX-9, was higher than that in uniform scaffolds or in gradual scaffolds without SOX-9. RT-PCR analysis also showed the highest expressions of Proteoglycan 4 (PRG4), Parathyroid hormone-related protein (PTHrP), type II collagen, aggrecan, SOX-9, and Hypoxia-inducible factor 1-alpha (HIF-1α) in the group of SOX-9 gradual scaffolds.

To better understand the single contributions of compression and shear forces in chondrogenic induction, Schatti et al. [27] analyzed these two stimuli, both alone and in combination. MSCs were seeded onto polyurethane scaffolds and underwent either compression at 1 Hz, in a strain amplitude of 10–20%, or oscillation of ±25◦ at 1 Hz, or a combination of both loads. Oscillation was imposed through the shifting of a ball perpendicularly to the scaffold axis. Superimposed compression was applied along the axis of the scaffold as well. The results suggested that stimulation by combined strains, instead of the application a single stimulus alone, is the best way to assure a chondrogenic phenotype in the absence of exogenous growth factors. The authors gave evidence of significant upregulation, in comparison to the control group, of the chondrogenesis markers (type II collagen, Aggrecan (AGG), Cartilage oligomeric matrix protein (COMP), SOX-9) only in the samples loaded with both compression and shear. Also, this last group was the only one in which type II collagen immunostaining was detected and that seemed to maintain a constant release of GAG in medium.

Cheng et al. [36] developed a novel construct made up by platelet-rich fibrin (PRF) membrane, which functions as a growth-factor-rich scaffold for bone-marrow-derived stem cells (BMSCs), for transplantation in cartilage defects. Flexibility of the neo-formed cartilage and differentiation of the stem cells were achieved through stimulation by hydrostatic pressure, in order to achieve boundaryless tissue consistency between the formed neocartilage and the damaged host cartilage in the temporomandibular joint (TMJ). TMJ offers an interesting environment for evaluating the integration of a cartilage construct within the damaged surface and to compare the behavioral differences of fibrocartilage in comparison to hyaline cartilage of the knee joint. In TMJ, the control of applied loads is difficult because of the impossibility to immobilize this joint, so the adaptation and responses of the construct to biophysical stimuli are more consistent and are extremely suitable to be studied as a model for cartilage regeneration approaches [37]. The hydrostatic pressure device, in this experiment, applied compression ranging from 90 to 150 kPa, revealing that proliferation and chondrogenic markers of the BMSC/PRF constructs were highest during the first days and gradually decreased at 6 days, suggesting that BMSCs could have limited chondrogenic capacity in relation to decreased growth factor release from the PRF. The authors concluded that pressure is an indispensable stimulus in order to promote cell proliferation, tissue regeneration, and repair mechanisms and to obtain a physiologic hierarchical and polar arrangement of the neoformed tissue.

### *4.1. Influence of Cell Distribution*

Gardner et al. [38] simulated the multiaxial mechanical loads that characterize the articular joint and observed the results derived from applying these forces on fibrin–poly(ester-urethane) scaffolds seeded with MSCs. The constructs were divided into three groups, represented in Figure 3: in the first group, the cells were evenly dispersed throughout the scaffold (Uniform); in the second group, the cells were asymmetrically disseminated within the matrix, forming a thin layer on the surface (Asymmetric); and in the third group, the cells were allowed to adhere only to the upper face of the scaffold (Surface Only). These different distributions were investigated in terms of matrix deposition in response to mechanical stress. The protocol employed 20 cycles of 10% compression, achieved by the raising and lowering of a ball onto the scaffold. The rotation on the ball generated shear friction of ±25◦ at 1 Hz. Histological and immunohistochemical analysis showed that there was an increase in glycosaminoglycans and type II collagen levels in the Asymmetric group in comparison to the other two groups. Besides this, the cells in the Surface Only group produced a small amount of matrix, suggesting a hypertrophic-like phenotype. In conclusion, the pattern of cell distribution within the scaffold is a critical parameter to take into consideration, and matrix deposition could be enhanced by considering the anisotropic properties of the materials.

Cell distribution can also be controlled by physiological mobilization of the cells, from their niches to different areas of the scaffold, employing biomechanical stimulation. Long-lasting regeneration of articular cartilage after surgical techniques like autologous chondrocyte implantation (ACI), mosaicplasty, and microfracture can be hampered by failure to attract progenitor cells, leading to the formation of fibrocartilage. In vitro loading compression provided by a bioreactor, such as 10% strain at 0.3 Hz frequency, applied intermittently for 24 hours, was shown to induce the mobilization of MSCs from a reservoir to an alginate scaffold located above it [39]. This experiment aimed to provide a model of a cartilage defect in the tibial plateau in order to evaluate the possible effect of biomechanical loading on cell recruitment from the subchondral bone. The mobilization of the stem cells from the reservoir required the supporting effect of laminin-521 (LN-521), as it is a basement

membrane protein which exerts a pivotal role in cell adhesion and migration mechanisms. The processes involved in antigravity migration within the scaffold are still the object of study and could imply extracellular signals between the cells and also physical fluidic movements induced by the stimulation of the bioreactor.

**Figure 3.** Different distributions of mesenchymal stem cells (MSCs) in fibrin–polyurethane composite scaffolds. Gardner et al., imposed a protocol of mechanical loads on scaffolds seeded with MSCs. The study comprised three types of distribution: cells evenly dispersed throughout the scaffold (Uniform), cells asymmetrically disseminated, forming a thin layer on the surface (Asymmetric), and cells allowed to adhere only to the upper face of the scaffold (Surface Only).

## *4.2. Osteogenic Involvement*

Another recent study by Carrol et al., presented the role of mechanical stimulation in MSC osteogenic differentiation [40]. This work explains about the ability of cyclic tensile strain (CTS) to regulate the initiation of MSC differentiation and, more specifically, their involvement in the endochondral pathway. MSCs embedded in fibrin hydrogels experienced uniaxial tensile deformation in a novel bioreactor system. The authors found out that CTS, in the absence of differentiation factors, can enhance the expression of tenogenic and osteogenic markers. The lack of evident chondrogenesis suggests that CTS could take part in directly initiating intramembranous ossification. When in the presence of chondrogenic growth factor (TGF-β3), instead, CTS induced increased proteoglycan and collagen production and enhanced upregulation of the markers of endochondral ossification (Bone morphogenetic protein 2 (BMP2), Runt-related transcription factor 2 (RUNX2), (Alkaline phosphatase) ALP, Osteopontin (OPN), Collagen Type X Alpha 1 Chain (COL10A1)). The authors concluded that CTS is an inducer of both endochondral and intramembranous ossification of stem cells, depending on the environment.

Endochondral ossification-based engineering techniques are promising strategies to provide regeneration of large defects. Hybrids of bone and cartilage tissue from induced pluripotent stem cells (iPSCs) have been achieved by the combined use of osteogenic and chondrogenic media in addition to mechanical stimulation, by means of shaking forces [41]. iPSCs offer the advantage of self-organizing in culture medium into cell aggregates, known as embryoid bodies (EB), without the support of scaffolds, providing a model to examine mechanisms of tissue differentiation and organ development. In this

experiment, the aggregates, when maintained in osteogenic medium culture, underwent osteogenic induction, reaching mineralization; when, instead, the aggregates were set first in osteogenic and later in chondrogenic medium, they were able to originate both tissues by expressing osteogenic and chondrogenic marker genes. The percentages of bone and cartilage composition seem to be subject to variations in culture periods, medium components, and shaking frequency, allowing the generation of easily manipulated osteochondral organoids.

## *4.3. Co-Cultures*

Co-cultures of articular chondrocytes (ACs) and MSCs have been proposed to overcome problems associated with the dedifferentiation of chondrocytes during in vitro expansion or to the tendency of MSCs to acquire hypertrophic features [42,43]. In co-cultures, the milieu created by the chondrocytes can, in fact, stimulate stem cell differentiation, which, in turn, acts as an enhancer of phenotype stability and the proliferation of chondrocytes by stimulating cell–cell adhesion mechanisms and by secreting several paracrine factors like growth factors and cytokines. Proper mechanical stress, in the form of cyclic sinusoidal dynamic tensile mechanical stimulation, can stimulate the co-culture by improving the deposition of ECM (GAGs, type II collagen) and the expression of regulatory factors (TGFβ, SOX9) and promoting the exchange of molecules between MSCs and chondrocytes [44].

Co-cultures offer the advantage of overcoming the tissue engineering challenge of maintaining a sufficient number of functional chondrocytes which will provide structural reliability with the right amount of ECM. Scaffolds embedded with ADSCs, subjected to cyclic compression in bioreactors, could benefit from the presence of chondrocytes, since they are able to release paracrine factors (i.e., TGF-β1 and IGF-1), guiding stem cells towards chondrogenic differentiation. This approach has been used to reduce the use of exogenous growth factors, which can instead be synthesized by a proper number of chondrocytes during in vitro expansion [35]. Even though similar production of type II collagen and GAGs was observed in single and co-cultures, the latter were shown to be able to suppress the expression of Col I, Col X, and Tumor necrosis factor alpha (TNF-α) markers [45].

#### *4.4. Computational Approach*

The spreading use of bioreactors has offered the possibility to observe, in a feasible and controlled way, the process of differentiation of stem cells under mechanical stresses, leading to expanded understanding of the mechanisms underpinning the relationship between physical stimulation and biological responses. The outcome of these kinds of stimulation is strongly dependent on the design of the machine, the type of mechanical forces, their intensity and frequency, and the times of stimulation, not to mention the variables related to the scaffolds and the cells. Therefore, the results coming from different studies lead to high heterogeneity and difficulties in providing a uniform consensus about the best protocol to induce processes of differentiation. For this reason, computational models can help, since they have been suggested and used to further investigate tissue engineering strategies. These approaches could be employed to characterize the mechanical stresses imposed by the bioreactors, refine scaffold geometry, and analyze physiobiological dynamics and cell behavior [46]. In silico experiments complement in vitro and in vivo analysis, addressing some complex questions which can be difficult to answer through more traditional approaches. Theoretical assumptions, simplifications in the conceptual framework, and the need to operate consistent validations could limit their use. Koh et al. [47] investigated the mechanisms of cartilage regeneration in osteochondral defects by using 3D medical imaging of the knee joint and analyzing the mechano-regulation processes underlying MSC differentiation. Two computational finite element (FE) models were employed to investigate the effects of physical stresses on cell regenerative mechanisms, as well as physiological processes like mitosis and cell death. The aim of the models was to provide predictions of the influence of different loading conditions on the whole system. The results indicated that simulation of a stance-phase gait cycle performed according to the ISO14,243-1 standard [48] induces more consistent cartilage regeneration than simulation of a vertical loading. This is because when the vertical loading was simulated, it

was predicted that endochondral ossification would sustain bone development and that hydrostatic pressure would induce the formation of fibrocartilage.

## **5. Conclusions**

Designing tissue engineering strategies for articular cartilage requires a thorough control of stem cell fate. The latter can be achieved through a more traditional approach using growth factors or, as suggested by the studies discussed above, by applying precise extrinsic mechanical loads, able to mimic the environment within the joint. Therefore, mechanical loading has been proposed as an alternative strategy to induce MSC chondrogenesis without the use of exogenous factors. Depending on the stimulus applied, promotion of specific tissue-related elements can be achieved. It can be concluded that shear stress and hydrostatic pressure, which exert their maximum effect on the superficial zone of the tissue, can increase type II collagen synthesis. Superimposed compression is, instead, the leading promoter of glycosaminoglycan production, since these components are directly involved in structural maintenance following perpendicular stresses. As a future prospective, functionalized bioengineering needs to take into account that de novo tissues should not lack the excellent organization of the cartilage that is critical for their role. A more in-depth study of the effects of different forces imposed on MSC cultures could represent a possible key to creating the original cartilage disposition with parallel collagen fibers in the superficial zone and perpendicular alignment in the thicker layer. Another aspect that should be investigated is the exact biomolecular mechanism by which the cells respond positively to friction and pressure. Tribology offers paramount principles to follow in order to understand and truly benefit the regenerative sciences.

**Author Contributions:** Conceptualization, S.R. and G.M.; writing—original draft preparation, S.R.; G.L.; writing—review and editing, M.A.S.; visualization, S.R.; supervision, G.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the University Research Project Grant (Triennial Research Plan 2016-2018), Department of Biomedical and Biotechnological Sciences (BIOMETEC), University of Catania, Italy.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## MDPI

St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Applied Sciences* Editorial Office E-mail: applsci@mdpi.com www.mdpi.com/journal/applsci

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18

www.mdpi.com

ISBN 978-3-0365-3940-9