Neonatal Achilles Tendon Microstructure is Negatively Impacted by Decorin and Biglycan Knockdown After Injury and During Development

Abstract

Interest in studying neonatal development and the improved healing response observed in neonates is increasing, with the goal of using this work to create better therapeutics for tendon injury. Decorin and biglycan are two small leucine-rich proteoglycans that play important roles in collagen fibrillogenesis to develop, maintain, and repair tendon structure. However, little is known about the roles of decorin and biglycan in early neonatal development and healing. The goal of this study was to determine the effects of decorin and biglycan knockdown on Achilles tendon structure and mechanics during neonatal development and recovery of these properties after injury of the neonatal tendon. We hypothesized that knockdown of decorin and biglycan would disrupt the neonatal tendon developmental process and produce tendons with impaired mechanical and structural properties. We found that knockdown of decorin and biglycan in an inducible, compound decorin/biglycan knockdown model, both during development and after injury, in neonatal mice produced tendons with reduced mechanical properties. Additionally, the collagen fibril microstructure resembled an immature tendon with a large population of small diameter fibrils and an absence of larger diameter fibrils. Overall, this study demonstrates the importance of decorin and biglycan in facilitating tendon growth and maturation during neonatal development.

Introduction

Tendons are a highly aligned and hierarchically organized collagen I-rich tissue that provide joint stability and transmit forces from muscle to bone. Tendon injuries present challenges to the individual by limiting daily activities and are increasingly common, resulting in a significant clinical burden [13, 14, 31]. Unfortunately, current therapeutics for tendon injury remain controversial with weak or inconclusive evidence for improving outcomes after injury [1]. Recently, neonatal mouse injury models have demonstrated an improved healing response in tendon, offering a new model to study for the future development of therapeutics. Unfortunately, little is known about neonatal tendon development and neonatal tendon healing.

Tendon development in the neonate mainly consists of collagen fibril growth and maturation, as the majority of collagen fibril assembly occurs during embryonic development [51]. Fibrillogenesis facilitates tendon growth and maturation, which is a multi-step process consisting of the production of collagen molecules to form fibril intermediates, linear growth of the fibril intermediates to produce longer fibrils, and finally lateral fusion of the fibrils to increase fibril diameter [51]. Early in development tendons have a narrow distribution of small diameter collagen fibrils [5]. As growth and maturation progresses, this distribution broadens as large diameter fibrils are accumulated. Modulation of fibrillogenesis during this phase of development is vital to ensure that the proper hierarchical structure of tendon is achieved since this process coincides with increases in ambulation, mechanical loading, and tendon mechanics [5, 38, 41].

The changes in tendon structure, mechanics, and composition in neonates are paralleled by the changes observed after tendon injury. For example, after injury, new collagen molecules are produced, which will lead to the production of larger diameter collagen fibrils via fibrillogenesis to establish the post-injury tendon structure. Due to this growth, similar changes to the collagen fibril diameter distribution are observed over time. Regulators of fibrillogenesis during tendon development, such as decorin and biglycan, are also upregulated after injury to prevent lateral fusion between immature collagen fibrils [50, 51]. Ultimately, the healing response does not replicate the original developmental processes that created the native tendon; healing produces scar tissue that is mechanically and structurally inferior. However, multiple neonatal Achilles injury models have demonstrated that when tendon injury occurs, a healing response that rapidly improves tendon structure and mechanics emerges [6, 25]. A neonatal Achilles tendon injury model can study the processes that regulate fibrillogenesis during tendon development and healing.

Decorin and biglycan are two class I small leucine-rich proteoglycans (SLRPs) that are key regulators of collagen fibrillogenesis and highly expressed in tendon. Decorin and biglycan share a similar structure, consisting of leucine-rich repeats connected to one or two glycosaminoglycan chains, respectively [27]. Expression of decorin and biglycan is high during neonatal development, but expression patterns vary with decorin being steadily expressed while biglycan peaks early and tapers off around postnatal day 7 [5, 50]. Both SLRPs have been shown to be key regulators of collagen fibrillogenesis during tendon development and healing, with deficiency of decorin and biglycan producing alterations to collagen fibril diameter and the presence of abnormally shaped fibrils [16, 32, 50, 51]. These changes to tendon structure impact tendon mechanics, producing changes in structural and material properties and viscoelastic mechanics depending on the model system [8, 15, 17–20, 45–47]. Decorin and biglycan also have roles in various signaling pathways including regulation of TGF-β activity, immune signaling through toll-like receptors, and the formation and maintenance of the tendon stem/progenitor cell (TSPC) niche [2, 3, 7, 10, 24, 39, 40]. These studies demonstrate the importance of decorin and biglycan in the development and maintenance of tendon structure and mechanics through direct regulation of fibrillogenesis and signaling across vital pathways. However, little has been done to explore the impact of these SLRPs on early neonatal tendon development when expression of decorin and biglycan is high, tendon growth and maturation is beginning, and the potential for an improved healing response is present.

Therefore, the objectives of this study were to determine the effects of decorin and biglycan knockdown on (1) the development of the neonatal Achilles tendon structure and mechanics and (2) the ability for the neonatal Achilles tendon to rapidly heal and recover biomechanical properties after injury. To achieve this, we used an inducible compound decorin/biglycan mouse model to knock down expression at two timepoints: on the day of birth (postnatal day 0) and at the time of injury (postnatal day 7). We hypothesized that knockdown of decorin and biglycan would (1) disrupt the neonatal development of tendon, resulting in decreased quasi-static and viscoelastic mechanical properties, altered collagen fiber realignment during loading, and altered collagen fibril diameter. We also hypothesized that (2) decorin and biglycan deficiency would impair the neonatal healing response by disrupting collagen fibril formation, producing injured tendons with impaired mechanics, reduced collagen fiber realignment, and altered collagen fibril diameter distributions with a large population of small diameter fibrils.

Materials and Methods

Animals

All methods utilized in this study were approved by the University of Pennsylvania Institutional Animal Care and Use Committee. The mice used in this study had a C57BL/6 background (Charles River, Wilmington, MA). Dcn^+/+/Bgn^+/+ (WT) were used as controls and compound Dcn^flox/flox/Bgn^flox/flox (I-Dcn^−/−/Bgn^−/−) mice with a tamoxifen (TM)-inducible Cre, (B6.129-Gt(ROSA)26Sortm1(cre/ERT2)Tyj/J, Jackson Labs, Bar Harbor, ME) were used as the experimental group. WT and I-Dcn ^−/−/Bgn^−/− mice were randomized into two groups: (1) uninjured mice with TM treatment on the day of birth (P0) that were sacrificed at postnatal day 7 (P7) (WT-P7 and I-Dcn ^−/−/Bgn^−/−-P7, n = 16/group) and (2) mice that received TM treatment at P7, prior to unilateral Achilles tendon injury, before being sacrificed 10-days post-injury at P17 (WT-P17 and I-Dcn^−/−/Bgn^−/−-P17, n = 20/group). All pups were administered 0.1 mg TM/1 g body weight (T5648, Sigma, St. Louis, MO) suspended in corn oil (C8267, Sigma). Pups that received TM at P0 were administered TM via intragastric injection, while pups treated at P7 were administered TM via intraperitoneal injection before surgery; a 29G needle was used for all pups. Each P7 mouse Achilles tendon was randomly assigned to an assay, with no two limbs from the same mouse being assigned to the same assay. P17 mice were randomly assigned to an assay, and each of the limbs (1 injured, 1 uninjured) were used in the same assay. Before tamoxifen treatment, the litter and the dam were placed in a separate cage, away from other mice in the colony, to avoid unintentional Cre activation.

Neonatal Achilles Tendon Surgery

After TM injection at P7, WT-P17 and I-Dcn^−/−/Bgn^−/−-P17 mice underwent surgery for a unilateral, full thickness, partial width (approximately 50%) Achilles tendon injury. Prior to surgery, animals were randomized and pain management was provided with a subcutaneous injection of Buprenorphine Sustained-Release. The animals were initially anesthetized using 3% isoflurane; isoflurane was maintained at 1.5% throughout surgery with a constant oxygen flow rate of 1 L/min. A 3 mm incision was made in the skin medial to the right Achilles tendon, then a rubber backing was placed beneath the center of the tendon to provide support. A 0.3 mm biopsy punch (RBP-03, Robbins Instruments, Houston, TX) was used to create a full thickness, partial width central defect in the Achilles tendon, then the skin was closed using a 6.0 prolene suture. Mice returned to normal cage activity and postoperative Buprenorphine Sustained-Release was given 3 days after surgery.

Gene Expression Analysis

Immediately after sacrifice, mouse Achilles tendons were dissected and stored in RNAlater (Thermo-Fisher, Waltham, MA) at −20 °C to stabilize the RNA (n = 3–5/group). Gene expression was measured using a Fluidigm 96.96 Dynamic Array^™ IFC (Fluidigm, San Francisco, CA) after samples underwent preparation consisting of RNA isolation, reverse transcription, and pre-amplification as previously described [33]. The Achilles tendons were prepared for RNA isolation by transferring the tissue to TRIzol (ThermoFisher) followed by disruption using a pestle and mortar and vortexing. Direct-zol RNA Microprep kits (Zymo Research, Irvine, CA) were used to isolate the RNA, then the RNA was reverse-transcribed using a High-Capacity cDNA RT kit (ThermoFisher). After reverse transcription, the cDNA for all targets, except for Col1a1, was pre-amplified for 15 cycles using Taqman Gene Expression Assays (Fluidigm), then loaded into a Fluidigm 96.96 Dynamic Array^™ IFC (Fluidigm) at the University of Pennsylvania Molecular Profiling Facility. Expression was measured across 92 genes; 18S, Rps17, and Abl1 were used as housekeeper genes. ΔCt values were determined by subtracting each Ct from the average Ct value of the housekeeper genes. ΔΔCt values were calculated by subtracting ΔCt values from the average WT ΔCt value.

Viscoelastic Biomechanical Testing

Prior to biomechanical testing, Achilles tendons (n = 8–12/group) were prepared as described [5]. Each sample was prepared under a dissection microscope to isolate the Achilles tendon and calcaneus and remove any excess tissue. After dissection, the cross-sectional area of the tendon was measured with a custom device using a laser and translational stages [43]. Cross-sectional area was measured every 0.5 mm along the length of the tendon to obtain 6–8 measurements per sample, then the average cross-sectional area was obtained from these measurements using custom Matlab software (Matlab2021a, Mathworks, Inc., Natick, MA). Verhoeff’s stain was used to apply stain lines at 1 mm intervals, beginning at the calcaneal insertion, for optical strain tracking. Sandpaper was glued to the calcaneal and myotendinous ends of the Achilles tendon, then secured using custom grips, to set a gauge length of 2.5 mm for mice sacrificed at P7 or 3 mm for mice sacrificed at P17. Next, the grips were secured using a custom grip holder to ensure that the sample would remain unloaded until it was mounted for mechanical testing [5].

The Achilles tendon was mounted in 1× phosphate-buffered saline bath at 37 °C within a tensile testing system (Instron 5848, Instron Corp., Norwood, MA) integrated with a custom cross-polarized light setup [30]. After the tendon was secured, it was preloaded to 0.01 N and an image was taken for gauge length measurements, then the viscoelastic mechanical testing protocol was initiated. The protocol consisted of a stress relaxation followed by a series of frequency sweeps and finished with a ramp to failure. During the ramp to failure a series of image maps were taken for optical strain tracking and collagen fiber realignment analysis; each image map consisted of 18 images obtained over 5 seconds, with 10 seconds separating each image map [30]. Collagen fiber realignment analysis was determined at 1%, 5%, 10%, 15%, and 20% strain using custom MatLab software (Matlab2021a) and outputted as circular variance ratio. Circular variance describes the distribution of collagen fiber realignment along the surface of the tendon and has an inverse relationship with collagen fiber realignment. The circular variance ratio is defined as the circular variance at a discrete strain level normalized by the circular variance of the tendon at gauge length. Force and displacement data obtained from the Instron during the ramp-to-failure were used to calculate toe and linear stiffness. The linear modulus was determined by using custom MatLab (MatLab2021a) software to analyze the images obtained throughout the linear region of the ramp-to-failure by tracking the stain lines on the tendon.

Transmission Electron Microscopy

Achilles tendons (n = 4/group) were analyzed using transmission electron microscopy (TEM). Samples were fixed in situ as described [11, 12, 45]. Post-staining with uranyless followed by 1% phosphotungstic acid, pH 3.2, was utilized for contrast enhancement. Cross sectional images within the midsubstance of the Achilles tendons were examined at 80 kV using a JEOL 1010 transmission electron microscope. Images were digitally captured at an instrument magnification of 60,000× using an 2 k × 2 k AMT CCD camera. Collagen fibril diameter was measured along the minor axis of the fibril using a custom program (Matlab2021a). Fibril diameter analyses were completed using blinded images from the center of the tendon. All fibrils within a predetermined region of interest (ROI) on the digitized image were analyzed. Non-overlapping ROIs were selected based on fibril orientation (i.e., cross section) and absence of cells.

Histology Preparation and Staining

Mouse hindlimbs were disarticulated at the hip, and the Achilles tendon was positioned perpendicular to the foot and placed in 4% paraformaldehyde at 4 °C for 2 days. After fixation, the limbs were transferred to a 30% sucrose solution at 4 °C for 24 h for cryopreservation then embedded in OCT media. 8 μm sagittal sections were collected throughout the tendon using Cryofilm Type IIC (Section Lab, Inc., Japan) [49], affixed to glass slides with 1.5% chitosan in 0.25% acetic acid, then sections were dried at 4 °C. Sections were stained with either hematoxylin and eosin or toluidine blue. P17 uninjured and injured sections were mounted with Hoechst 333424 (ThermoFisher) as a nuclear marker. Toluidine blue staining was quantified by determining the percentage of the tissue positive for metachromatic staining. Image analysis was performed using Fiji while blinded [48].

Statistics

Mechanical properties and histological parameters were compared using Student t-tests. A two-way repeated measures ANOVA w/Bonferroni post-hoc analysis was used to compare fiber realignment data using genotype and strain as factors. Collagen fibril diameter distributions were compared using the Kolmogorov-Smirnov test, and F tests were used to compare collagen fibril diameter variance. Principal component analysis (PCA) was performed on ΔΔCt values via scikit-learn [42]. Data preprocessing for PCA and gene expression analysis was conducted using NumPy and pandas [22, 36]. Matplotlib was used for PCA data visualization [26]. Principal component (PC) scores and ΔCt values were compared using Student t-tests. Significance was set at p ≤ 0.05.

Results

Decorin and Biglycan were Effectively Knocked Down in P7 and P17 Mice

P7 I-Dcn^−/−/Bgn^−/− Achilles tendons had a 4.25-fold decrease in decorin expression and a 4.54-fold decrease in biglycan expression relative to WT (Fig. 1a, b). No changes to fibromodulin, lumican, keratocan, or asporin expression were found in P7 tendons (Fig. 1c–f). Uninjured P17 I-Dcn^−/−/Bgn^−/− decorin expression decreased by 2.40-fold and biglycan expression decreased by 2.44-fold, while injured P17 I-Dcn^−/−/Bgn^−/− tendon expression decreased by 2.54-fold and 3.53-fold for decorin and biglycan, respectively, compared to P17 WT tendons (Fig. 1g, h). No changes in fibromodulin, lumican, keratocan, or asporin were present in injured P17 I-Dcn^−/−/Bgn^−/− tendons (Fig. 1i–l). Asporin and keratocan expression was increased in uninjured P17 I-Dcn^−/−/Bgn^−/− tendons (Fig. 1k, l).

Fig. 1.

SLRP expression in P7 and P17 Achilles Tendons. Dcn and Bgn expression was measured in P7 and P17 tendons after tamoxifen treatment. Knockdown of Dcn and Bgn was confirmed in P7 I-Dcn^−/−/Bgn^−/− tendons (a, b). No compensatory changes in SLRP expression were observed at P7 (c–f), Dcn and Bgn knockdown was also confirmed in P17 I-Dcn^−/−/Bgn^−/− tendons that were injured and uninjured (g, h), while there were no changes to Fmod or Lum expression (i, j). However, increased expression of Kera (k) and Aspn l was present in uninjured I-Dcn^−/−/Bgn^−/− tendons. Data shown as average with standard deviation. Each symbol marks a unique data point, while a solid line indicates the zero-point

P7 PCA captured 95% of the variance in six PCs (PC1: 35.0%, PC2: 24.9%, PC3: 15.0%, PC4: 11.6%, PC5: 6.4%, PC6: 5.1%). Lower PC3 scores were observed in I-Dcn^−/−/Bgn^−/− tendons (Fig. S1 a, b). I-Dcn^−/−/Bgn^−/− tendons also showed increased Axin2 expression and reduced Col12a1 and Prg4 expression (Fig. S1 c–e). Additional P7 gene expression results are displayed in Supplemental Table 1.

PCA on P17 ΔΔCt values for uninjured tendons captured 95% of variance across six PCs, and across five PCs for injured tendons. Higher PC1 scores were observed in uninjured I-Dcn^−/−/Bgn^−/− tendons (Fig. S2 a, b), while injured I-Dcn^−/−/Bgn^−/− tendons showed lower PC1 scores (Fig. S2 c,d). Differential gene expression was observed in 42 genes in uninjured I-Dcn^−/−/Bgn^−/− tendons (Table S2) and 46 genes in injured I-Dcn^−/−/Bgn^−/− tendons (Table S3), with 19 and 23 unique genes in each group respectively.

Decorin and Biglycan Deficiency Produced Neonatal Tendons with Decreased Cross-Sectional Area and Decreased Stiffness

Knockdown of decorin and biglycan produced tendons with decreased cross-sectional area at P7 (Fig. 2a). During mechanical testing, it was found that decorin and biglycan deficiency led to tendons with decreased linear stiffness and toe stiffness compared to WT (Fig. 2b, c). There were no changes to transition strain or linear modulus (Fig. 2 ), and no changes to tendon relaxation, dynamic modulus, or phase shift (Fig. 2e–k).

Fig. 2.

P7 Achilles Tendon Quasistatic and Viscoelastic Mechanics. Achilles tendon Quasistatic mechanics were measured during the ramp-to-failure in the mechanical testing protocol. I-Dcn^−/−/Bgn^−/−-P7 cross-sectional area a was reduced compared to WT-P7. A trending decrease was observed in toe stiffness b while linear stiffness was decreased (c). Linear modulus was not affected after knockdown (d). Viscoelastic mechanics were obtained by performing a stress relaxation followed by a series of frequency sweeps. Knockdown of decorin and biglycan did affect stress relaxation (e), dynamic modulus (f–h), or phase shift i–k at 0.1, 1, or 5 Hz. Data shown as average with standard deviation. Each symbol marks a unique data point

Collagen Fiber Realignment Ended Prematurely In I-Dcn^−/−/Bgn^−/−-P7 Tendons

WT-P7 and I-Dcn^−/−/Bgn^−/−-P7 Achilles tendon collagen fiber realignment increased from 1 to 5%, 5 to 10%, and 10 to 15% strain during the ramp-to-failure at the end of the mechanical testing protocol (Fig. 3). However, only WT-P7 realignment increased between 15 to 20% strain, whereas collagen fiber realignment in Achilles tendons deficient in decorin and biglycan did not continue to realign at these higher strain levels.

Fig. 3.

P7 Achilles Tendon Collagen Fiber Realignment. Collagen fiber realignment was measured during the ramp-to-failure and calculated at 1%, 5%, 10%, 15%, and 20% strain during mechanical testing. WT-P7 collagen fiber realignment increased between 1%-5%, 5%-10%, 10%-15%, and 15%-20% strain. I-Dcn^−/−/Bgn^−/−-P7 fiber realignment increased between 1%-5%, 5%-10%, and 10%-15% strain. Data shown as average with standard deviation. Bars represent group data, indicating average values

I-Dcn^−/−/Bgn^−/−-P7 Tendons Displayed a Substantial Decrease in Collagen Fibril Diameter

Utilizing TEM for collagen fibril diameter analysis, I-Dcn^−/−/Bgn^−/−-P7 Achilles tendon collagen fibril distributions displayed changes in diameter compared to the WT-P7 tendons, as shown in the representative images (Fig. 4a, b). Both groups exhibited unimodal collagen fibril diameter distributions (Fig. 4c). Tendons deficient in decorin and biglycan revealed smaller diameter collagen fibrils (Fig. 4d), decreased collagen fibril diameter variance, and a negatively skewed distribution relative to WT tendons. I-Dcn^−/−/Bgn^−/−-P7 collagen fibril diameter was reduced at the 25th (43.66 nm versus 46.85 nm), 50th (53.23 nm versus 58.58 nm), and 75th (61.41 nm versus 72.03 nm) percentiles compared to WT-P7 (Fig. 4e).

Fig. 4.

P7 Achilles Tendon Collagen Fibril Diameter. Collagen fibril diameter was measured using transmission electron microscopy from the tendon midsubstance. Representative images are displayed for WT a and I-Dcn^−/−/Bgn^−/− b tendons. The collagen fibril diameter frequency distribution binned fibrils into 5 nm bins. The WT-P7 distribution shows an accumulation of large diameter fibrils, while I-Dcn^−/−/Bgn^−/− tendons show a large population of small diameter fibrils with an absence of larger diameter fibrils (c). Comparisons between the groups confirmed a decrease in fibril diameter in the I-Dcn^−/−/Bgn^−/− tendons (d). The box and whisker plot captures 1–99th percentiles, with values outside of this range shown as individual points. Breakdown of each group by quartile showed that I-Dcn^−/−/Bgn^−/− collagen fibril diameter was lower at the first, second, and third quartiles (e)

Post-Injury Mechanics Were Impaired in Decorin and Biglycan Deficient Achilles

Following knockdown of decorin and biglycan, there were no alterations to cross-sectional area (Fig. 5a, b). No changes to stiffness were seen in the uninjured group (Fig. 5c), while injured I-Dcn^−/−/Bgn^−/− tendons displayed a decrease in toe stiffness (Fig. 5d). Similarly, no changes to maximum force were observed in uninjured tendons (Fig 5e), though maximum force in injured I-Dcn^−/−/Bgn^−/− tendons decreased (Fig. 5f). No changes were present in P17 linear stiffness (Fig 5g, h) or maximum stress (Fig. 5i, j). Linear modulus decreased in both uninjured (Fig. 5k) and injured (Fig. 5l) I-Dcn^−/−/Bgn^−/− tendons at P17. Lastly, no changes were observed in viscoelastic mechanics after knockdown of decorin and biglycan when measuring stress relaxation, dynamic modulus, or phase shift (Fig. 5m–r).

Fig. 5.

Quasistatic and Viscoelastic Mechanics of P17 Achilles Tendons. Quasi-static mechanics were obtained at the end of the mechanical testing protocol during the ramp-to-failure. No changes to cross-sectional area were present (a, b). Uninjured tendons showed no changes to toe stiffness (c), while toe stiffness decreased in I-Dcn^−/−/Bgn^−/− tendons (d). Max force was also unchanged in uninjured tendons (e), though max force decreased after injury (f). There were no changes to linear stiffness (g, h) or max stress (I, j). Deficiency of decorin and biglycan resulted in decreased linear modulus in uninjured and injured tendons (k, l). Viscoelastic mechanics were obtained by performing a stress relaxation followed by a series of frequency sweeps. No changes in viscoelastic mechanics were found in uninjured or injured I-Dcn^−/−/Bgn^−/− tendons (m–r). Data shown as average with standard deviation. Each symbol marks a unique data point (a–n), while bars represent group data (o–r)

Minimal Collagen Realignment Occurred in Injured I-Dcn^−/−/Bgn^−/−-P17 Achilles

Uninjured WT-17 and I-Dcn^−/−/Bgn^−/−-P17 tendon collagen realignment occurred from 1 to 5% strain and 5 to 10% strain. WT-17 tendon realignment peaked between 10 and 15% strain, with no additional realignment occurring between 15 and 20% strain (Fig. 6). In contrast, I-Dcn^−/−/Bgn^−/−-P17 tendon realignment did not increase between 10 to 15% strain, with increased realignment observed between 15 and 20% strain. Collagen fiber realignment of injured WT-17 and I-Dcn^−/−/Bgn^−/−-P17 tendons increased between 1 and 5% strain. The WT-17 group displayed continued realignment between 5 and 10% strain and 10 and 15% strain, whereas I-Dcn^−/−/Bgn^−/−-P17 realignment slowed considerably with no changes detected beyond 5% strain (Fig. 6).

Fig. 6.

P17 Achilles Tendon Collagen Fiber Realignment. The P17 tendon collagen fiber realignment was observed during the ramp-to-failure during mechanical testing and calculated at 1%, 5%, 10%, 15%, and 20% strains. Injured WT-P17 tendons between 1–5%, 5–10%, and 10–15% strains, while injured I-Dcn^−/−/Bgn^−/−-P17 tendons realigned between 1–5% strain. Realignment occurred in uninjured WT-P17 tendons between 1–5%, 5–10%, and 10–15% strains. Uninjured I-Dcn^−/−/Bgn^−/−-P17 tendon realignment occurred between 1–5%, 5–10%, and 15–20% strains. Data shown as average with standard deviation. Bars represent group data, indicating average values

Uninjured I-Dcn^−/−/Bgn^−/− Achilles Tendons Displayed a Decreased Collagen Fibril Diameter

Representative images shown in Figure 7 provide a view of the uninjured WT (Fig. 7a) and I-Dcn^−/−/Bgn^−/− (Fig. 7b) collagen fibrils at P17. TEM analysis showed that the collagen fibril diameter distributions in the uninjured WT and I-Dcn^−/−/Bgn^−/− tendons were broadly dispersed, with a decrease in collagen fibril diameter observed after knockdown of decorin and biglycan (Fig. 7c, d). This is supported by breakdown of the distributions into quartiles, where the I-Dcn^−/−/Bgn^−/− group displayed smaller diameter fibrils at the 25th (53.17 nm vs 55.55 nm), 50th (71.76 nm vs 75.77 nm) and 75th (88.73 nm vs 94.64 nm) percentiles (Fig. 7e).

Fig. 7.

Uninjured P17 Tendon Collagen Fibril Diameter. Collagen fibril diameter was measured using transmission electron microscopy from the tendon midsubstance. Representative images are displayed for WT a and I-Dcn^−/−/Bgn^−/− b tendons. The uninjured collagen fibril diameter frequency distributions showed similarly broad distributions c with a shift towards smaller diameter fibrils in the I-Dcn^−/−/Bgn^−/−-P17 group (d). The box and whisker plot captures 1–99th percentiles, with values outside of this range shown as individual points. Breakdown of the distributions by quartile revealed differences between the two groups are primarily among the large diameter fibrils (e). Data shown as average with standard deviation

Injury Impacts the Collagen Fibril Structure in I-Dcn^−/−/Bgn^−/− Achilles Tendons

The impact of decorin and biglycan knockdown on post-injury tendon microstructure is displayed in the representative images (Fig. 8a, b) where I-Dcn^−/−/Bgn^−/− Achilles tendon collagen fibril diameter was decreased relative to WT (Fig. 8c, d). The I-Dcn^−/−/Bgn^−/− fibril population was highly concentrated, with over half of the population having diameters between 30 and 45 nm. This reduction in fibril diameter was present across the 25th, 50th, and 75th percentiles (Fig. 8e).

Fig. 8.

Injured P17 Tendon Collagen Fibril Diameter. Collagen fibril diameter was measured using transmission electron microscopy from the tendon midsubstance. Representative images are displayed for WT a and I-Dcn^−/−/Bgn^−/− b tendons. The uninjured collagen fibril diameter frequency distributions showed similarly broad distributions c with a shift towards smaller diameter fibrils in the I-Dcn^−/−/Bgn^−/−-P17 group (d). The box and whisker plot captures 1–99th percentiles, with values outside of this range shown as individual points. Breakdown of the distributions by quartile revealed differences between the two groups are primarily among the large diameter fibrils (e). Data shown as average with standard deviation

Histological Analyses Revealed No Changes After Knockdown of Decorin and Biglycan

No differences in cellularity or nuclear aspect ratio were observed in P7 or P17 I-Dcn^−/−/Bgn^−/− tendons. Toluidine blue staining was also unchanged in decorin and biglycan deficient tendons at P7 and P17 timepoints (Fig. S3).

Discussion

The objectives of this study were to explore the effects of decorin and biglycan deficiency on (1) neonatal development of tendon structural and mechanical properties and (2) the healing response of Achilles tendon in the neonate. To achieve these objectives, we used a TM-inducible compound Dcn/Bgn mouse knockdown model. This model allowed the mice to undergo normal development up until TM treatment, thereby isolating the effects of decorin and biglycan knockdown to the desired experimental timeline. We hypothesized (1) that knockdown of decorin and biglycan at postnatal day 0 would disrupt neonatal tendon development and produce tendons with decreased quasi-static and viscoelastic mechanical properties, altered collagen fiber realignment, and an altered tendon collagen fibril diameter distribution compared to WT Achilles tendons. Additionally, we hypothesized (2) that knockdown of decorin and biglycan at postnatal day 7, concurrent with the Achilles tendon injury, would impair the neonatal healing response by disrupting collagen fibril formation, leading to tendons with impaired mechanics, reduced collagen fiber realignment, and altered collagen fibril diameter distributions, characterized by a high population of small diameter fibrils. In support of our first hypothesis, I-Dcn^−/−/Bgn^−/−-P7 Achilles tendons revealed reduced CSA and stiffness, early cessation of collagen fiber realignment during tensile loading, and TEM analysis revealed a unimodal collagen fibril distribution composed predominantly of small to medium diameter fibrils. The results indicated that the second hypothesis was also supported; knockdown of decorin and biglycan disrupted the neonatal healing response. During mechanical testing, I-Dcn^−/−/Bgn^−/−-P17 injured Achilles tendons had a decreased maximum force, toe stiffness, and linear modulus, altered collagen fiber realignment dynamics, and a collagen fibril distribution with a disproportionately large number of small fibrils compared to WT-P17 tendons.

Gene Expression and Effects of Decorin and Biglycan Knockdown

Effective knockdown of Dcn and Bgn was demonstrated for all three I-Dcn^flox/flox/Bgn^flox/flox experimental groups. Expression of Fmod, Lum, Kera, and Aspn was measured due to the structural and functional similarities of these SLRPs with Dcn and Bgn [23, 28, 34]. No changes to SLRP expression were found in P7 or injured P17 tendons after knockdown of Dcn and Bgn. Upregulation of Kera and Aspn was present in uninjured P17 I-Dcn^−/−/Bgn^−/− tendons, indicating that the effects from the knockdown model may be partially masked in this group due to the ability of Kera and Aspn to regulate fibrillogenesis via collagen binding and the overlap in signaling pathways [9, 29, 44]. Overall, the neonatal I-Dcn^flox/flox/Bgn^flox/flox mouse model demonstrated knockdown of the target genes with minimal upregulation of class I and II SLRPs.

Implications for Neonatal Tendon Microstructure and Collagen Fiber Realignment

TEM analysis found that decorin and biglycan deficiency at P7 produced many small diameter fibrils with an absence of larger diameter fibrils, which should be accumulating by P7 [4]. I-Dcn^−/−/Bgn^−/−-P17 tendons revealed similar outcomes, although the deviation from WT-P17 was most striking in the injured tendons. Knockdown of decorin and biglycan also reduced collagen fibril diameter variance across all groups, which is indicative of a reduction in fibril diameter heterogeneity and supports previous observations that decorin and biglycan are regulators of lateral growth of collagen fibrils. The impact of decorin and biglycan knockdown on tendon microstructure were reflected in the collagen fiber realignment. WT-P7 tendons showed a steady increase in collagen realignment throughout the toe and linear regions, while I-Dcn^−/−/Bgn^−/−-P7 realignment occurred rapidly in the toe region, before peaking in the linear region at 15% strain. Injured I-Dcn^−/−/Bgn^−/−-P17 tendon realignment also showed an early peak at 5% strain before slowing at higher strains while having a high population of small collagen fibrils present. The immature collagen fibril structures observed in the I-Dcn^−/−/Bgn^−/−-P7 and injured I-Dcn^−/−/Bgn^−/−-P17 tendons are the likely early peaks in collagen fiber realignment compared to their respective WT controls, since developmental age and underlying ECM fibril structure is thought to influence tendon realignment [37].

Impacts of Decorin and Biglycan Knockdown on Neonatal Tendon Mechanics

Additionally, the effect of decorin and biglycan knockdown on the underlying collagen fibril structure impacted tendon mechanics, which were negatively influenced regardless of age or state of injury. I-Dcn^−/−/Bgn^−/−-P7 tendon cross-sectional area was reduced when the neonate is undergoing rapid growth and increasing ambulation [5, 35]. Linear stiffness and toe stiffness were also reduced, while linear modulus was unchanged. While neonates grow rapidly at P7, ambulation is still relatively low, so the effect of the immature collagen fibril structure may have a more significant impact on material properties with increased mechanical loading and a longer developmental timeline. I-Dcn^−/−/Bgn^−/−-P17 linear modulus was decreased in both injured and uninjured tendons, indicating that the reduced modulus in the injured tendon is partially due to knockdown of decorin and biglycan and not necessarily indicative of an impaired healing response. Interestingly, viscoelastic mechanics were unaffected by decorin and biglycan knockdown regardless of age or injury. Previous studies using SLRP knockdown models commonly reported viscoelastic mechanics being more sensitive to the impact of SLRP deficiency on tendon mechanics than quasi-static mechanics [17, 20, 21]. Previous studies using the I-Dcn^−/−/Bgn^−/− model in mature and aged mice showed few changes to tendon quasi-static mechanics and, especially in aged mice, drastically altered viscoelastic mechanics, making the lack of changes to viscoelastic mechanics in this study particularly surprising [9, 45].

Limitations and Future Work

This study is not without limitations. Gene expression was measured 7 and 10 days after knockdown. While expression data suggests sufficient knockdown of Dcn and Bgn occurred, protein levels were not measured, though gene expression was considered most important. Additionally, both Dcn and Bgn were knocked down in the experimental group, thus this study is unable to differentiate between the roles of decorin and biglycan during neonatal development. Future studies will utilize single knockdown models to explore the individual roles of decorin and biglycan during neonatal tendon development. No data was collected measuring gait or ground reaction forces throughout neonatal development. Measuring these outcomes would help determine if changes to tendon structure and mechanics manifest in joint function and could be included in future work. Finally, additional timepoints to study the effects on development and healing on a longer timeline would provide further insight on the impacts of decorin and biglycan knockdown in the neonate.

Conclusion

Overall, this study demonstrated that decorin and biglycan deficiency negatively impacted neonatal development and the neonatal healing response. P7 and injured neonatal tendons had large populations of small diameter fibrils, indicating that decorin and biglycan are vital to fibrillogenesis in these contexts. The small diameter fibrils likely caused the negatively impacted collagen fiber realignment and quasistatic mechanics. This study provided a better understanding of the roles of decorin and biglycan in the Achilles tendon after injury and during neonatal development. Establishing a better understanding of these processes can help inform future therapeutics for tendon injury and aging.