Introduction

Tendinopathy is the most common musculoskeletal complaint for which patients seek medical attention1,2,3,4,5. It is prevalent in occupational and athletic settings, afflicting 25% of the adult population and accounting for 30–50% of sport injuries2. The incidence is rising in line with major risk factors, including repetitive overuse due to mechanically demanding work and sports activities, age, obesity and diabetes5. Thus, tendon injuries pose a substantial socioeconomic burden and an escalating challenge to healthcare services, with the annual health expenditure exceeding €145 billion1,2,3,4,5.

Due to its low cellularity, poor vascularity and bradytrophic metabolism, a tendon’s response to injury is inefficient and results in a fibrovascular scar that never attains the gross, histological or mechanical characteristics of a normal tendon, leading to high re-injury rates and chronic tendinopathy2,4. Current treatment fails to restore the functional properties of injured tendons, significantly decreasing mobility, productivity and quality of life5,6,7. In an increasingly active and ageing population, the long-term morbidity of tendon disease represents a growing unmet clinical need. However, research efforts are hampered by the low yield of primary tenocytes, their limited proliferation rate and lifespan in vitro and the corresponding lack of relevant in vitro models8,11. Tenocytes in culture undergo senescence and a phenotype drift, losing their fusiform morphology and the expression of typical lineage markers such as Collagen1 (Col1), Tenomodulin (Tnmd), Scleraxis (Scx) and Tenascin C (Tnc)10,12,13,14 at passage 3–510,11,12,14, which restricts the number and quality of cells available for research and thereby experimental design. The resulting need to sacrifice high numbers of animals to obtain sufficient tenocytes for research poses ethical problems. Donor variability further increases the number of animals needed to differentiate inter-individual variability from the effects caused by experimental conditions15. Efforts to immortalize tenocytes by delivering the oncogene SV40 large T antigen to murine tenocytes using a piggyBac transposon system successfully prolonged the lifespan of the cells to passage 30 but drastically impaired tendon marker expression (Col1, Col3, Scx, Tnmd and Tnc)10. In contrast, three-dimensional culture in collagen gels improves proliferation rates, tendon marker expression and cellular yield compared to 2D culture but does not overcome the problem of senescence10. Thus, to date, in vitro tendon research possibilities are limited, and better research tools are urgently required to advance the development of satisfactory treatment options.

The cell cycle regulators p16Ink4a/Arf and p14ARF (p19ARF in mice) are encoded by the Ink4a/Arf locus. p16Ink4a/Arf is a cyclin-dependent kinase inhibitor that prevents G1–S phase progression. p14ARF influences responses to a wide range of cellular stress signals by interfering with p53–mediated transcription. Both proteins play an essential role in cell growth, survival and senescence and act as tumour suppressors16,17,18,19,20. Targeted disruption of the Ink4a/Arf locus in mice generates a cancer-prone phenotype16,17,18,19,20,21. Mutations in the Ink4a/Arf locus in humans are frequent and found in multiple malignancies, including hematopoietic tumours, melanoma, carcinoma and sarcoma16. The expression of p16Ink4a/Arf and p19ARF increases markedly with ageing22,23 and contributes to the decline of the replicative potential of stem cells22. In the ageing Achilles tendon, upregulation of Ink4a/Arf locus leads to exhaustion of the tendon stem cells pool in size and function24,25,26 and may contribute to the decreased healing potential following injury in aged tendons. Deletion of the Ink4a/Arf locus has been shown to immortalize various cell types, including MEFs (mouse embryonic fibroblasts)21,27.

This study establishes an immortalized tenocyte cell line from Achilles tenocytes isolated from Ink4a/Arf deficient mice (Ink4a/Arf−/−). The immortalized tenocytes behave like their wild-type primary counterparts in monolayer and 3D culture but in contrast to wild-type primary cells preserve their proliferation potential and tenogenic properties even during long-term culture and after freeze–thaw cycles, thus providing a valuable in vitro alternative to animal models for tendon research.

Results

Ink4a/Arf−/− tenocytes proliferate consistently while maintaining a tenogenic morphology

Western blot analysis demonstrated p16INK4a in wt tenocytes but a complete loss of p16INK4a protein expression in the Ink4a/Arf−/− cells, confirming the successful deletion of the locus in the Ink4a/Arf−/− tenocytes (Fig. 1a, Supplementary Fig. S1a).

Figure 1
figure 1

Characterization of Ink4a/Arf−/− compared to wild-type (wt) tenocytes (n = 3 different biological replicates, ** p-value < 0.01, *** p-value < 0.001, for mean, s.d. and p-values see Table 1) (a) Western blot of p16 and α-Tubulin (as loading control) in Ink4a/Arf−/− and wt cells confirmed the successful deletion of p16Ink4a/Arf in the Ink4a/Arf−/− cells. (b) The proliferation capacity, illustrated as the number of cells obtained over four days, of intermediate and high passages Ink4a/Arf−/− tenocytes was significantly higher than of wt tenocytes. (c) Senescence, illustrated as the level of β-galactosidase activity, of Ink4a/Arf−/− tenocytes was significantly lower than in wt tenocytes (a.u.—arbitrary units). (d) DNA content/Aneuploidy, intracellular DAPI staining of wt low passage (3 independent donors), Ink4a/Arf−/− low, intermediate and high passage (3 independent cell lines).

Full size image
Figure 2
figure 2

Morphology, cell area and circularity of Ink4a/Arf−/− cells in comparison to wild-type (wt) cells in different passages. **p-value < 0.01, ***p-value < 0.001. a–e wt tenocytes in a low passage (a, passage 2–5) and Ink4a/Arf−/− tenocytes in a low (b, passage 2–5), intermediate (c, passage 14–17) and high (d, passage 35–44) passage showed similar tenogenic morphology while in wt tenocytes in passage 6 (e) phenotypic drift was already apparent. Micrographs were obtained using phase contrast with a 40 × objective (scale bar = 1000 µm). (f) Cell area (μm2) of the wt cells and the Ink4a/Arf−/− cells in different passages. Wt cells in passage 6 showed significantly higher cell area indicative for dedifferentiation. (g): Circularity (0–1 range) of the wt cells and the Ink4a/Arf−/− cells in different passages. Circularity in wt cells in passage 6 was significantly closer to 1 than the rest of the groups suggesting a rounder appearance indicative for dedifferentiation.

Full size image

Ink4a/Arf−/− tenocytes maintained their proliferation potential up to passage 44, the highest passage included in this study. The proliferation rates of Ink4a/Arf−/− cells at the intermediate and higher passage remained consistently high. Ink4a/Arf−/− tenocytes proliferated even faster than wt controls and Ink4a/Arf−/− cells at a low passage (Fig. 1b, Table 1).

Table 1 Summary of proliferation, senescence (β-galactosidase activity), flow cytometry and qPCR (gene expression) analyses performed in wild-type (wt) and Ink4a/Arf−/− tenocytes.
Full size table

Senescence, measured by ß-Galactosidase activity, remained significantly lower in Ink4a/Arf−/− tenocytes than in wt tenocytes even in high passages (p-value < 0.01, Fig. 1c, Table 1). Aneuploidy, examined by flow cytometry of intracellular DAPI staining, was not apparent in Ink4a/Arf−/− tenocytes and DNA content remained constant in high passages (Fig. 1d, Supplementary Fig. S1b). The higher proliferation of Ink4a/Arf−/− tenocytes was visible by a higher fraction of cells in the S phase of the cell cycle (chromosome duplication, 4n; Fig. 1d, Supplementary Fig S1b). Furthermore, low, intermediate, and high passage Ink4a/Arf−/− cells maintained their fusiform morphology and adherence to the plastic plate surface, comparable to wt cells (Fig. 2a–d). In contrast, wt cells changed their morphology at passage six and drifted towards a rounder and larger appearance with less defined margins and a flattened shape indicative for dedifferentiation (Fig. 2e–g).

Figure 3
figure 3

Tenocytes from Ink4a/Arf−/− mice maintain their viability, low senescence and tenogenic morphology following freeze–thaw cycles. *p-value < 0.05, **p-value < 0.01. (a) Flow cytometry analysis of Annexin-V staining for low passage wild-type (wt) cells, Ink4a/Arf−/− cells at an intermediate passage and Ink4a/Arf−/− cells at an intermediate passage following a freeze and thaw cycle (n = 3 biological replicates). Wt cells showed significantly lower viability than Ink4a/Arf−/− cells at an intermediate passage with and without a freeze–thaw cycle. (b) Senescence, illustrated as the level of β-galactosidase activity, was significantly higher in wt than in Ink4a/Arf−/− cells at intermediate passage and Ink4a/Arf−/− cells at intermediate passage following freeze and thaw cycle (n = 3 different biological replicates) (a.u.—arbitrary units). (c) Ink4a/Arf−/− cells at intermediate passage following freeze and thaw cycle showed tenogenic morphology when imaged using phase contrast with a 40 × objective (scale bar = 1000 µm).

Full size image

Ink4a/Arf−/− tenocytes maintain cell viability and tenogenic features following freeze–thaw cycles

Viability of Ink4a/Arf−/− tenocytes was not affected by the freeze–thaw cycle (75.6% ± 4 live cells following a freeze–thaw cycle versus 74.2% ± 9 before freezing, p-value = 0.98) and remained significantly higher than the viability of wt cells (34.2% ± 10, p-value < 0.01) (Fig. 3a). The freeze–thaw cycle did not induce senescence in Ink4a/Arf−/− cells (115 a.u. ± 32 before freezing, 162 arbitrary units (a.u.) ± 134 after a freeze–thaw cycle, p-value = 0.87), as shown by the consistently lower β-galactosidase activity of Ink4a/Arf−/− compared to wt cells (528 a.u. ± 123, p-value < 0.05) (Fig. 3b). Cell morphology remained unchanged, tenocytes maintained their fusiform appearance without any evidence of a rounded or flattened phenotype indicating dedifferentiation (Fig. 3c).

Figure 4
figure 4

Tenocytes from Ink4a/Arf−/− mice in low, intermediate, and high passage exhibit tenogenic characteristics comparable to tenocytes from wild-type (wt) mice in low passage. (a) Box & whiskers plots (mean indicated by the horizontal line and whiskers indicating min–max range) illustrating the qPCR data of tendon markers Col1, Col3, Col5, Scx, Tnmd and Tnc in wt cells and Ink4a/Arf−/− cells from different passages (n = 3 different biological replicates). For mean normalized mRNA levels, s.d. and p-values see Table 1. (b) Representative histograms of flow cytometry analysis of CD29, CD105, CD90 and CD44 expression in wt (low passage) and Ink4a/Arf−/− cells (in low, intermediate, and high passages) revealed similar levels of surface markers confirming their tenogenic phenotype (n = 3 different biological replicates).

Full size image

Ink4a/Arf−/− tenocytes maintain tenogenic characteristics in culture

The expression of tenocyte markers (collagen type 1 (Col1), Col3, Col5, Scleraxis (Scx), Tenomodulin (Tnmd) and Tenascin-C (Tnc)) expression of Ink4a/Arf−/− cells at low, intermediate, and high passage remained stable at all time points investigated and equivalent to wt cells (Fig. 4a, Table 1). Also the levels of the surface markers CD29, CD90, CD105 and CD4428 measured by flow cytometry were comparable between Ink4a/Arf−/− and wt cells (Fig. 4b, Table 1). CD29 and CD44 expression was high (> 95%) in both cell populations and remained high in all Ink4a/Arf−/− passages investigated. CD90 expression was high (> 90%) in wt tenocytes and Ink4a/Arf−/− cells in early and intermediate passages but decreasing in higher passages (p = 0.009, Supplementary Fig S2, Table 1). CD105 was expressed in only a small number (38%) of low passage cells of both wt and Ink4a/Arf−/− genotype and in fewer cells in the intermediate (p-value = 0.36) and high (p-value = 0.16) passages of Ink4a/Arf−/− cells (Supplementary Fig S2, Table 1).

Ink4a/Arf−/− tenocytes produce 3D tendon-like constructs

To test the functional properties, Ink4a/Arf−/− tenocytes of all passages were evaluated in 3D tendon-like constructs. Wild-type cells served as control and verified the comparable architecture (Fig. 5a).

Figure 5
figure 5

Tenocytes from Ink4a/Arf−/− mice can produce 3D tendon-like constructs similar to wild type (wt) cells. (a) The macroscopic appearance of the silicone plate with the tendon-like constructs before harvesting (left) with a close-up of the tendons in their wells (right). (b)–(e) Histology of H&E stained tendon-like constructs obtained from wild-type (wt) cells in a low passage (b) and Ink4a/Arf−/− cells in low (c), intermediate (d) and high (e) passages (top row: scale bar 400 µm, bottom row: scale bar 100 µm). (f) 3D tendon-like constructs of Ink4a/Arf−/− tenocytes in an intermediate passage following freeze and thaw cycle showed analogous cell and fiber arrangement to 3D tendon-like constructs of wt or Ink4a/Arf−/− cells, which had not undergone a freeze–thaw cycle (left micrograph: scale bar = 400 µm, right micrograph: scale bar = 100 µm). (g) qPCR analysis of tendon markers- Col1, Col3, Col5, Scx, Tnmd and Tnc presented as fold change (FC) in gene expression between 3 and 2D, for wt and Ink4a/Arf−/− cells individually. The dotted line in each graph represents a FC of 1 (no difference between 2 and 3D gene expression). The expression in Col1, Col3, Col5 and Tnc, but not in Tnmd and Scx, was significantly higher in 3D than 2D culture for both wt and Ink4a/Arf−/− cells (for mean normalized mRNA expression and p-values, see Table 2).

Full size image

Histological analysis confirmed the tendon structure of the 3D constructs characterized by a longitudinal arrangement of extracellular matrix (ECM) with cells scattered therein arranged parallel to the fibers irrespective of the phenotype and passage number (Fig. 5b–e). This remained unchanged after freeze/thaw cycles (Fig. 5f). As expected, and in line with the literature, tendon markers Col1, Col3, Col5, and Tnc were significantly higher expressed in the 3D tendon-like constructs than in 2D monolayer culture (Fig. 5g, Table 2). The expression of Tnmd and Scx was comparable between 2D- and 3D-cultured cells. No difference was noted between the wt and the cell line in expression of the tendon markers in the 3D tendon-like constructs.

Table 2 qPCR analysis of tendon markers expression in wild-type (wt) and Ink4a/Arf−/− tenocytes in 2D and 3D culture.
Full size table

Discussion

The current study successfully established an immortalized Ink4a/Arf−/− tenocyte cell line, which combines a long lifespan with a stable tenogenic phenotype and high proliferation capacity. The tenocytes isolated from the Achilles tendons of Ink4a/Arf−/− mice maintain their proliferation potential and tenogenic fusiform morphology even in long-term culture (passage 44 compared to a maximum of 5 passages for wt tenocytes) and express significantly lower levels of the senescence marker ß-gal compared to wt tenocytes. Ink4a/Arf−/− cells express tendon markers and create 3D tendon-like constructs analogous to their wt counterparts. Based on these results, we conclude that deletion of the Ink4a/Arf locus does not interfere with tenogenic features. Furthermore, Ink4a/Arf−/− cells also preserve their viability and morphology following freezing and thawing, facilitating their use for research purposes.

Immortalization, or the ability to avoid senescence, is crucial in overcoming the finite in vitro proliferative potential of primary mammalian cells, the so-called Hayflick limit. Immortalization is frequently accompanied by the deregulation of tumour suppressors and thus bears the potential risk of facilitating malignant transformation29. The Ink4a/Arf locus is one of the most frequently mutated sites in human cancer16,18. In the current study, no tumorigenesis was observed. As the mice included in this study were young, further studies evaluating the effect of p16Ink4a/Arf deletion on tumorigenesis in old mice are warranted.

Ink4a/Arf−/− tenocytes do not undergo senescence and continue to proliferate for many passages (passage 44) with a steady expression of tendon markers. In contrast to previous studies reporting an early increase in proliferation rates in Ink4a/Arf−/− embryonic fibroblasts19,20, Ink4a/Arf−/− tenocytes proliferated at a comparable rate in low passages as wt tenocytes. Growth rates increased significantly in intermediate and high passages while the cells preserved their tenogenic characteristics.

In addition to the improved lifespan of the tenocytes cell line, the Ink4a/Arf−/− cells maintained their tenogenic morphology, tendon marker expression and the ability to form 3D tendon-like constructs comparable to primary wt cells. Identifying and validating a tenocyte phenotype is hindered by the lack of tendon-specific markers as the tendon ECM genes such as Col1, Col3, Col5 and Tnc are also expressed in other musculoskeletal tissues14,30. Therefore, a panel of markers, including not only ECM genes but also genes associated with tendon development, such as Scx and Tnmd, are required to specify tenogenic characteristics14,30. In the current study, common tendon markers (Col1, Col3, Col5, Scx, Tnmd and Tnc) were expressed in Ink4a/Arf−/− cells of all passages equivalent to wt cells. This finding is important since previous tenocyte cell lines showed a significant reduction in tendon markers gene expression compared to wt tenocytes11. Flow cytometry is less frequently used for tendon characterization since the markers are more established for progenitor cells (e.g. mesenchymal stem cells or tendon progenitor cells) or fibroblasts in general28,31,32,33,34. However, although mesenchymal stem cells and tenocytes are different cell types, similar fibroblastic surface marker expression has previously been reported34,35 and adds another valuable tool to define a tenogenic phenotype. In our study, primary wt cells and Ink4a/Arf−/− tenocytes in all passages highly (> 95%) expressed the markers CD29 and CD44, but CD90 and CD105 expression decreased with passaging. This finding supports a previous report of murine mesenchymal stem cells expressing CD105 to a similar percentage with decreasing CD105 expression during passaging36. Regarding the loss of CD90 expression during passaging, the literature is equivocal with some studies reporting a decrease of CD90 expression in stem cells with passaging, while others did not observe the same phenomenon24,34,37,38.

In recent years, 3D models have gained popularity since they provide a higher resemblance of the morphological and functional properties of the tissue as compared to monolayer cell culture10,39,40,41. In the current study, the creation of 3D tendon-like constructs was essential to assess the function of the cells as tenocytes align to the direction of the anchor points, contract and produce ECM. The Ink4a/Arf−/− cells formed those constructs in all passages within the same time frame (10 days) as wt tenocytes, demonstrating their tenogenic function. Col1, Col3, Col5 and Tnc gene expressions revealed the tendon markers expression is significantly higher in 3D compared to 2D monolayer. This tendency was not shown in Scx and Tnmd, which revealed a similar, very low expression in both 2D and 3D. The higher expression of tendon markers in the 3D tendon-like constructs emphasizes the importance of studying tenocytes in systems resembling the natural environment, which supports the full functionality of the cells.

In summary, we successfully established an immortalized tenocyte cell line isolated from the Achilles tendon of Ink4a/Arf depleted mice, which maintains tenocytes’ morphological and functional features. Compared to wt cells, the Ink4a/Arf−/− cells have higher proliferation rates and lower senescence, while the expression of tendon markers, morphology, and ability to create 3D tendon-like constructs remain similar to wt tenocytes. In addition, Ink4a/Arf−/− tenocytes can reach high passage numbers and withstand freeze–thaw cycles while maintaining tenogenic characteristics. Combining a long lifespan in vitro with a stable tenogenic phenotype, the Ink4a/Arf−/− tenocyte cell line will enable more extended experiments and significantly facilitate research into the highly prevalent, disabling tendon diseases. Notably, the phenotypically stable Ink4a/Arf−/− tendon cell line will also limit the number of animals needed for each experiment and enable the development of in vitro alternatives to animal trials.

Methods

Mice

C57BL/6N (wild-type = wt, n = 3) and Ink4a/Arf-deficient (Ink4a/Arf−/−, n = 3)20 female mice were housed under Specific-pathogen-free (SPF) conditions at the University of Veterinary Medicine Vienna, Austria. All experiments were conducted in 8- to 12-week-old mice. Mice were placed under general anesthesia with isoflurane inhalation and euthanized using cervical dislocation for reasons unrelated to this study. Both Achilles tendons were harvested immediately following euthanasia.

Experiments were approved by the institutional ethics and animal welfare committee and the national authority (license BMWF-68.205/0218-II/3b/2012), in accordance with ARRIVE guidelines. All methods were performed in accordance with the relevant guidelines and regulations implemented at the University of Veterinary Medicine Vienna, the Institutional Ethics Committee (“Ethics and Animal Welfare Committee”) of the University of Veterinary Medicine Vienna.

Tenocytes isolation and cell culture

The Achilles tendons of both hind limbs were harvested, washed with phosphate-buffered saline (PBS), and sectioned into 5 mm pieces following removal of the paratenon. Tenocytes were isolated by enzymatic digestion using 3 mg/mL Collagenase Type II (Gibco Lifetechnologies, Vienna, Austria) in culture media (minimal essential medium) supplemented with 10% fetal bovine serum (FBS), 1% L-Glutamine, 100 U/mL penicillin and 0.1 mg/mL Streptomycin for 6 h. After washing with PBS (centrifugation for 5 min at 400 g), cells were cultivated in culture media at 37 °C and 5% CO2 until the desired passage and number of cells were obtained.

Experiments were performed for the wt cells in a low passage (passage 2–5) and the Ink4a/Arf−/− cells in 3 different passages—low (passage 2–5), intermediate (passage 14–17) and high (passage 35–44).

All experiments were carried out with biological triplicates (the same three wt and three Ink4a/Arf−/− mice for all experiments).

For freeze–thaw cycle assessment, Ink4a/Arf−/− cells were frozen in 10% DMSO and 90% FBS at an intermediate passage and were used for experiments 2–3 passages following thawing.

Western blot analysis

Western blotting (WB) was performed using standard protocols42. Cells from Ink4a/Arf−/− mice at intermediate passage were compared to wt cells. Total protein extraction was done with Laemmli lysis buffer, and protein concentration was determined using a BCA assay (Thermo scientific No. 23225). Proteins were separated on an SDS–polyacrylamide electrophoresis (PAGE) gel and transferred to PVDF membranes in an overnight wet transfer. Membranes were incubated overnight at 4 °C with primary antibodies, followed by 1 h room temperature incubation with secondary antibodies using monoclonal rabbit anti-mouse p16INKa (Abcam, ab211542; dilution 1: 2,000), monoclonal chicken anti-mouse α-Tubulin (DM1A, Sigma-Aldrich, CP06; dilution 1: 1000). Chemiluminescence detection was performed using the Bio-Rad ChemiDoc Imaging systems.

Proliferation assay

Cells were plated in 96-well plates (3000 cells/well). The proliferation rate was measured via DNA fluorescence using the CyQuant assay (Invitrogen, C35012) 24, 48, 72 and 96 h after seeding.

Cellular senescence/β-galactosidase activity

Cells were plated in 96-well plates (3000 cells/well). Cellular senescence was quantified by measuring senescence-associated β-galactosidase (SA-β-gal) activity according to the manufacturer’s instructions (SA-β-gal activity, Cell Biolabs, CBA-231) 48 h after seeding.

Morphology

Cells were imaged at 60–70% confluency using the EVOS FL Auto imaging system in phase contrast with a 40 × objective (ThermoFisher Scientific, AMEP4680).

Cell area and circularity were measured as previously described43. Shortly, the boundaries of individual cells were manually traced on the images using ImageJ 1.53t software to calculate the cell area and the perimeter of each cell. Circularity was defined as C = 4π(A/P2), where P is the perimeter, and A is the cell area. Circularity value of 1 indicates a perfect circle while value of 0 indicates an elliptical shape.

Flow cytometry

Cells (1 × 105) were stained with fluorophore-conjugated antibodies for surface markers CD29, CD44, CD90 and CD10528 (CD29-PE (HMb1-1; eBioscience), CD44-PE (IM7; eBioscience), CD90-FITC (30-H12; BioLegend), CD105-PE (MJ7/18; eBioscience)) diluted 1:200 in PBS with 2% FBS and incubated for 30 min at 4 °C. Cells were washed twice with PBS and measured on a BD FACS Canto II. To detect apoptotic cells, tenocytes were resuspended in 1 × Annexin-V binding buffer and stained with Annexin-V for 15 min at room temperature according to the manufacturer’s instructions (eBioscience, 88-8005-74). To analyze the DNA content, the respective cells were stained with DAPI (D9542, Sigma). Therefore, 1 × 105 cells were washed with PBS, resuspended in 2% paraformaldehyde in PBS and incubated for 10 min at 37 °C. After centrifugation, the supernatant was discarded and cells were permeabilized by 90% methanol in PBS with 2% FCS and 0.2% Tween-20 on ice for 30 min. Cells were washed twice and stained with 2 µg/ml DAPI in PBS with 2% FCS and 0.2% Tween-20 for 10 min on ice followed by washing twice. Data was acquired on a Beckman Coulter Cytoflex S. Flow cytometry data was analyzed by BD FACSDiva Software or FlowJo (v10.6.1).

RNA extraction and quantitative real time-polymerase chain reaction (qRT-PCR)

Tenocytes were seeded in 12-well plates (1 × 105 cells/well). After 24 h, RNA was isolated using Trizol and Chloroform in a ratio of 5–1. Total RNA was recovered by adding isopropanol and glycerol. The mixture was placed on ice and centrifuged for 60 min at 13,000 rpm. The total RNA pellet was washed twice with 75% ethanol and solubilized in RNase-free water. DNA remains were removed by a DNA removal kit (Life Technologies). 5 ng RNA was used for the qPCR reaction (qPCR One-Step Eva Green kit, Bio&amp; Sell, Feucht, Germany). The results were analyzed using the QuantStudio (Invitrogen, Thermo Fischer Scientific) software. The mRNA levels were calculated by first normalizing using the BestKeeper Index (Housekeeping genes used- GAPDH, 36B4, Rps18) and then using the 2^ct∆ method. All primers (Supplementary Table S3) were designed using the Primer3 software.

3D tendon-like constructs

Tendon-like constructs (1 cm long) were produced by providing tenocytes suspended in Collagen type 1 gel with anchor points embedded into the wells (well dimension- 10 mm long, 3 mm wide and 2 mm deep) of a 31-well silicon plate. The 31-well plate was custom-made using silicon (Sylgard 184 Silicone elastomer Dow Corning) and a 3D-printed template (Supplementary Fig S4). After autoclaving, the silicone plate was placed in a 100/20 mm cell culture dish, and each well was coated with a biocompatible polymer (Lipidure, AMS biotechnology) to prevent attachment of the cells to the silicone. Tenocytes were suspended in a 60% α-MEM/ 40% collagen (PureCol EZ Gel solution, Sigma-Aldrich) mixture kept on ice. Sixty μl of the solution containing 2.5 × 104 cells were added to each well. Following 2 h of incubation at 37 °C and 5% CO2 to allow the gel to solidify, 2 pins (Austerlitz insect’s pins, 0.2 mm) were inserted into each tendon at the edge of each well and culture medium was added. Culture medium was replaced every 3 days. 3D tendon-like constructs formed gradually over a few days, during which the cells and collagen arranged longitudinally. By day 10, the 3D tendon-like constructs were visible and stable. The 3D tendon-like constructs were analyzed by qRT-PCR and histology.

For qRT-PCR, 31 of the 3D tendon-like constructs were pooled and RNA was extracted as described earlier. Analysis was done as a fold change (FC) of the 3D tendon-like construct expression relative to the average 2D (monolayer) expression. A fold change higher than 1 will mean higher expression of the marker in the 3D tendon-like construct in comparison to the 2D monolayer. The low, intermediate, and high passages of the Ink4a/Arf−/− groups were pooled together into one group named Ink4a/Arf−/−.

Histology

For histology, the 3D tendon-like constructs were washed with PBS and fixed with formalin solution while still in the silicone plate. Following formalin fixation, the 3D tendon-like constructs were removed, dried, and embedded in paraffin. Paraffin-embedded 2-µm sections were stained with hematoxylin and eosin G. Micrographs of the 3D tendon-like constructs were taken using the EVOS FL Auto imaging system in colour phase with a 100 × and 400 × objective (ThermoFisher Scientific, AMEP4680).

Statistical analysis

Statistical analyses were performed using Graph Pad Prism software (version 8.4.3).

Continuous variables were expressed as mean ± standard deviation (s.d.), and categorical variables were expressed as percentages. Differences between wt and Ink4a/Arf−/− cells and between the different passages of Ink4a/Arf−/− cells were analyzed using a t-test or ANOVA with Tukey's multiple comparisons test. A p-value < 0.05 was considered significant.