Achilles and tail tendons of perlecan exon 3 null heparan sulphate deficient mice display surprising improvement in tendon tensile properties and altered collagen fibril organisation compared to C57BL/6 wild type mice

Tissues

Hspg2^{Δ3−∕Δ3−} homozygous mouse breeding pairs backcrossed into a C57BL/6 background for 12 generations were kindly supplied by Dr. R Soinninen, University of Oulu BioCentre, (Oulu, Finland). WT C57BL/6 mice were obtained from Jackson Laboratories (Bar Harbor, ME, USA). All mice were caged in groups (n = 2–5 mice per 500 cm² cage floor space) and received acidified water and complete pelleted food ad libitum. All cages were individually ventilated with filter lids, sterilised Aspen chip bedding, environmental enrichment (tissues, house) and maintained at 20–22 °C, 50–60% humidity, with a 12 h light-dark cycle regimen. Only male mice were used for these studies.

Overview of the analyses undertaken in this study

1. Genotyping of WT and Perlecan exon 3 null mice.

2. Weight of WT and Perlecan exon 3 null mice up to 20 weeks of age.

3. HS contents of purified perlecan samples measured by ELISA.

4. Material properties (UTS/TM) of tail and Achilles tendons over 3–12 weeks of age.

5. Compositional analyses of the GAG and hydroxyproline contents of tail tendons in WT and Perlecan exon 3 null mice up to 12 weeks of age.

6. Material properties of control non operated Achilles tendons over 2–8 weeks of age and the recovery of material properties (UTS/TM) of tenotomised Achilles tendons in WT and Perlecan exon 3 null mice over 2–8 weeks post operatively.

7. Collagen I, versican, biglycan, lumican, decorin, perlecan, LTBP-1 and 2, fibrillin-1, elastin gene expression by 3–12 week old WT and Perlecan exon 3 null tail tenocytes.

8. Response of WT and Perlecan exon 3 null mouse tail tenocyte monolayer cultures to FGF-2 (1–100 ng/ml): Mmp2, Mmp3, Mmp13, Adamts4, Timp2, Timp3 gene expression.

9. Transmission electron microscopy of collagen fibril cross-sectional areas in tail and Achilles tendon in WT and Perlecan exon 3 null mice at 3 and 12 weeks of age.

Genotyping of Hspg2^{Δ3−∕Δ3−} mice

Genomic DNA was extracted from WT and Hspg2^{Δ3−∕Δ3−} mouse tail tips using commercial kits (Qiagen). Specific regions of the mouse perlecan gene were amplified by PCR using genotyping primers recognising intron 2 of mouse Hspg2 (GTA GGG ACA CTT GTC ATC CT), exon 3 (CTG CCA AGG CCA TCT GCA AG) and Hspg2^{Δ3−∕Δ3−} (AGG AGT AGA AGG TGG CGC GAA GG). The PCR products were identified by electrophoretic separation on 2% w/v agarose gels (Fig. S1A).

Weights of WT and Perlecan exon 3 null mice

Mouse weights were measured in each genotype over 10–20 weeks of age (Fig. 1B).

Isolation and identification of perlecan from skeletal muscle

Muscle from the hind limbs of two WT and two Hspg2 exon 3 null mice were finely minced and extracted with 6M urea 50 mM Tris–HCl pH 7.2 (15 ml/g tissue) for 48 h at 4°C. Perlecan was isolated using Resource Q anion exchange FPLC and electrophoresed on pre-poured 3–8% PAG Tris-acetate gradient gels, blotted to nitrocellulose and perlecan identified using MAb H300 (Santa Cruz Biotechnology, Dallas, TX, USA) (Fig. S1B). The GAG side chains of these samples were analysed by ELISA using MAb 10-E-4 and 3-G-10. Selected samples were pre-digested with Heparitinase III to generate Δ-HS stub epitopes reactive with MAb 3-G-10.

Biomechanical assessment of murine tail and Achilles tendons

Tail tendons from 3, 6 and 12 week old mice, (n = 6–8 at each age point), were dissected from underlying bone and connective tissue within 1 h of death. Mouse tails contain two dorsal and two ventral tendons. Ventral tendons were used exclusively in this study. These were dissected out approximately 5 mm from the base of the tail. Segments of 20–30 mm were used for biomechanical testing. A single ventral tendon was used for biomechanical testing per animal. The cross-sectional area of the tendon was measured using a custom micrometer device (Choi et al., 2016). Achilles tendons attached to the calcaneous and gastrocnemius muscles proximally were also collected. Biomechanical testing was undertaken in an Instron 8874 servo-hydraulic material testing apparatus. Tail tendons were marked with two reference points using Alcian blue and anchored in custom-built brass clamps. The distance between the clamps was 10 mm. With Achilles tendons the calcaneous was anchored in the lower clamp, and gastrocnemius muscle attached to the upper clamp. Tendons were loaded to failure at a rate of 1 mm/second (10% strain), real-time videos of each test were recorded at 100 frames per second using a high-speed camera mounted perpendicular to the tendon (Marlin F/145B camera; Allied Vision Technology, Newburyport, MA, USA). The videos of tendon deformation were analysed using LabView software (National Instruments, Austin, TX, USA) and the data normalised to tendon cross-sectional area to calculate stress. Elastic modulus was calculated from the gradient of the linear region on the stress–strain curve.

Tendon compositional analyses

Finely-diced tail tendons were papain digested and sulfated GAG determined using 1,9-dimethylmethylene blue, bovine tracheal CS was used as standard (Farndale, Buttle & Barrett, 1986). Aliquots of the tissues were also hydrolysed in 6M HCl for hydroxyproline determinations by the dimethylaminobenzaldehyde procedure (Stegemann & Stalder, 1967).

Achilles tendon tenotomy

Under general anaesthesia (2% isofluorane) Achilles tendons of 12 week old mice from each genotype were sharply transected mid-way between the calcaneal attachment and muscle leaving the plantaris tendon intact. The skin incision was closed with a subcutaneous Vicryl 8/0 suture and sealed with cyanoacrylate tissue glue. Mice were returned to their pre-operative social groups post-tenotemy. Groups of mice (n = 10 for each time point) were sacrificed at 2, 4 or 8 weeks post operatively (PO) and Achilles material properties measured.

Gene expression in tail tendons

Mouse tail tendons from 3, 6, and 12 week old WT and Hspg2^{Δ3−∕Δ3−} mice were pooled (n = 6 at each age point) to provide ∼50 mg wet weight of tissue for each RNA isolation. Tendons were snap frozen in liquid nitrogen, freeze-shattered using a Mikro dismembranator (B. Braun Biotech International, Melsungen, Germany) and total RNA extracted using Trizol (Invitrogen, Mulgrave, VIC, Australia), purified using Qiagen RNeasy columns (Qiagen, Chadstone Centre, VIC, Australia) and quantified by NanoDrop (ThermoFisher Scientific, Scoresby, VIC, Australia). RNA (1 µg) from each sample was reverse transcribed using Omniscript Reverse Transcription Kit (Qiagen) with random pentadecamers (50 ng/ml; Sigma-Genosys, Castle Hill, NSW, Australia) and RNase inhibitor (10 U per reaction, Bioline, Sydney, NSW, Australia). The cDNA was subjected to qRT-PCR in a Rotorgene 6000 (Qiagen) using Immomix (Bioline, Sydney, NSW, Australia), SYBR Green I (Cambrex Bioscience, Rockland, ME, USA), and 0.3 µM validated murine-specific primers. Relative copy numbers for genes of interest were determined using a standard curve generated from pooled cDNA normalised to Gapdh. PCR primer specificity was confirmed by sequencing (SUPAMAC, Sydney University). Genes, primers, and annealing temperatures are listed in Table 1.

Tenocyte monolayer cultures stimulated with FGF-2

Three week old tail tendons from six mice of each genotype were macerated and cultured in 2 ml DMEM/10% FBS/2 mM L-glutamine under an atmosphere of 5% CO₂, with media changes every 3-4 days. After 2 weeks the tissue was removed and the attached cells detached with trypsin-EDTA and sub-cultured in fresh media. Passage 3 cells were cryopreserved (10⁷ cells/ml, 0.5 ml aliquots) in 10% v/v DMSO, 20% v/v FBS in DMEM. Tenocytes were re-seeded in 6-well plates at 2 × 10⁵ cells per well for 48 h. The cultures were washed three times in serum-free DMEM, and incubated in DMEM/1% v/v FBS containing 0, 1, 10 or 100 ng/ml FGF-2 (PeproTech Inc, Rocky Hill, NJ, USA) for 24 h. Total RNA was extracted using Trizol, 1 µg RNA was reverse-transcribed then qRT-PCR undertaken for Mmp2, Mmp3, Mmp13, Timp1, Timp3 and Adamts4.

Transmission electron microscopy of tail and Achilles tendon

Three tendons from 3- and 12-week old WT and Hspg2 ^{Δ3−∕Δ3−} mice were washed in 0.1 M sodium cacodylate buffer containing 3 mM CaCl₂, 100 mM sucrose pH 7.4 and fixed in 2.5% v/v glutaraldehyde/0.5% v/v paraformaldehyde for 30 min at room temperature, 4 °C for 24 h, followed by storage in 70% v/v ethanol. The fixed tissues were trimmed and post-fixed in 2% w/v OsO₄ in 0.1M cacodylate buffer for 1–2 h at 4 °C followed by dehydration in serial graded ethanol washes (25%, 50%, 75%, 95%, 100%, 100%, all v/v). The tissues were infiltrated with Spurrs resin/ethanol (1:1) overnight then with two overnight infiltrations of 100% resin and polymerised at 60 °C for 48 h. Ultra-thin transverse tissue sections (70 nm) were cut using an Ultracut T microtome and transferred to copper grids (200 mesh). The specimens were stained/contrasted for 10 min with 2 % w/v uranyl acetate and Reynold’s lead citrate (1.33 g lead nitrate, 1.76 g sodium citrate dihydrate, 5 ml 1 M NaOH, in 50 ml H₂O final total volume). The specimens were examined in a JEOL1400 transmission electron microscope at 120 kV at 25,000 × magnification. The images were analysed using ImageJ (public domain Java-based image processing software developed by NIH) to determine fibril diameters. Three separate regions of each specimen were photographed and at least 400 fibrils were measured in each image. When the fibril had a non-circular configuration the diameter across the minimum axis was measured. The frequency distribution of the collagen fibril diameters was calculated as a percentage of the total fibril numbers measured.

Statistical analyses

Groups of paired parametric data from mechanical measurements, hydroxyproline and sGAG compositional data and fibril diameters measured by TEM were compared by unpaired Students-t test for differences between age and genotypes. Non-Gaussian data from qRT-PCR analyses were analysed using Kruskal-Wallis then Mann–Whitney U ranked tests. The alpha level was set at 0.05.

Genotyping of mouse strains

Figure 1A depicts the replacement of perlecan exon-3 with a pGK-neo cassette in the Hspg2^{Δ3−∕Δ3−} mice. Hspg2^{Δ3−∕Δ3−} mice were fertile and litters were of expected size. They had no gross abnormalities or difference in appearance compared to WT mice at birth. By 3 weeks of age, the previously reported micropthalia in Hspg2^{Δ3−∕Δ3−} animals was evident. Age-matched Hspg2^{Δ3−∕Δ3−} mouse body weights were less than corresponding WT mice 10 to 18 weeks of age (Fig. 1B). Although smaller, Hspg2^{Δ3−∕Δ3−} mice had similar skeletal proportions to WT mice and no apparent musculoskeletal abnormalities. Mutant mice were more docile when handled but no other behavioural abnormalities were noted other than small leaky eyes.

Perlecan isolated from C57BL/6 Wild Type and Hspg2 exon 3 null mice

Perlecan isolated from skeletal muscle demonstrated a core protein of ∼420 kDa for WT perlecan while mutant perlecan had a molecular weight of ∼300–390 kDa (Fig. S1B). ELISA analysis demonstrated that perlecan from C57BL/6 Wild Type mice contained HS chains while Hspg2 exon 3 null perlecan did not (Fig. 1C).

Tendon material properties and biochemical composition

Ultimate tensile stress (UTS) (Figs. 2A, 2C) and tensile modulus (TM) (Figs. 2B, 2D) measurements of tail (Figs. 2A, 2B) and Achilles (Figs. 2C, 2D) 3–12 week old tendons demonstrated there were no significant difference in 3-week old tendons. UTS and TM values for tail and Achilles tendons increased with maturation (Figs. 2A–2D). Six to twelve week old tail tendons from the perlecan exon 3 null mice displayed moderately greater UTS and TM values compared to WT tail tendons (P < 0.05) (Figs. 2A, 2B). Tail tendon sGAG levels underwent a maturation-dependent decline in Hspg2 ^{Δ3−∕Δ3−} mice (Figs. 2E, 2F), with 50% reduction in GAG content at 12 weeks compared to 3 weeks (3 >6 >12 week), and lower GAG content in Hspg2^{Δ3−∕Δ3−} mice whereas hydroxyproline contents did not change significantly with age or genotype.

Achilles tendon tenotomy

With tissue maturation the non-operated contralateral perlecan exon 3 null Achilles tendons displayed moderately higher UTS and TM values by 20 weeks of age compared to the WT tendons (P < 0.05). Following tenotomy, WT and perlecan exon 3 null Achilles tendons partially recovered their UTS and TM material properties over an 8 week recovery period reaching approximately 50% of the UTS/TM values of the non-operated contralateral tendons at 20 weeks of age (Figs. 3A, 3C). There was a moderately increased trend in the rate of recovery of UTS/TM in the WT mice compared to the perlecan exon 3 null mice following tenotomy. The UTS and TM values for the WT tendons however were more variable than the perlecan exon 3 null tendons. Statistics on the mean data at each time point for each genotype indicated this trend was only moderately elevated in the WT Achilles tendons with a P < 0.05 (Figs. 3B, 3D).

Tail tendon gene expression

qRT-PCR of selected ECM genes in mouse tail tendons, (Col1a1, Vcn, Bgn, Dcn), demonstrated a maturation-dependent decline in gene expression over 3 to 12 weeks in both genotypes (Fig. 4). Relative gene expression levels for perlecan core protein (Hspg2) and the elastin micro-fibril associated proteins Ltbp1, Fbn1, and Eln were significantly lower in Hspg2^{Δ3−∕Δ3−} mice compared to age-matched WT mice. Ltbp2 gene expression displayed an increase with ageing in both genotypes.

Tendon outgrowth cell responses to FGF-2

Mmp2, Mmp3, Mmp13 and Adamts4 expression were significantly higher in basal Hspg2^{Δ3−∕Δ3−} mice compared to WT tenocyte cultures (0 ng/ml FGF-2). Mmp2 expression in Hspg2^{Δ3−∕Δ3−} cultures remained significantly higher than WT at all doses of FGF-2 examined. Mmp 3 and Mmp13 expression increased dose-dependently in both genotypes, and this response was greater in Hspg2 ^{Δ3−∕Δ3−} tenocytes at high doses of FGF-2 (Mmp 3 33–36 fold versus 50 fold and Mmp13 134–192 fold versus 226–248 fold at 10 and 100 ng/ml in WT and Hspg2^Δ3−∕Δ3) (Fig. 5). Adamts4 gene expression was significantly decreased by increasing concentrations of FGF-2 treatment in Hspg2^{Δ3−∕Δ3−} tenocyte cultures but still remained significantly greater than WT cultures at all doses. FGF-2 up-regulated Timp1 gene expression, less so in Hspg2^{Δ3−∕Δ3−} compared to WT cultures, favoring a pro-catabolic phenotype in the mutant mice. Expression of Timp3, the naturally occurring inhibitor of the ADAMTS (A Disintigrin And Metalloprotease with Thrombospondin motifs) enzymes, while equivalent in basal culture, was decreased to a greater extent by FGF-2 in Hspg2^{Δ3−∕Δ3−} tenocyte cultures at the highest dose.

Measurement of Tail and Achilles tendon collagen fibril diameters by transmission electron microscopy

Figure 6 shows the changes in collagen fibril diameter of tail and Achilles tendon measured from TEM images of 3 and 12 week old mouse tendons. In 3 week old mice there were small differences between genotypes in mean fibril diameter (±SD) in tail (WT = 202 ± 60 nm, Hspg2^{Δ3−∕Δ3−} = 193 ± 46 nm; not significant) and Achilles (WT = 160 ± 44 nm, Hspg2^{Δ3−∕Δ3−} = 139 ± 37 nm; P < 0.001) tendons (Figs. 6A, 6B). This was reflected in minor differences in frequency distributions of collagen fibril diameters in the two genotypes in these immature mice (Figs. 6C, 6D). An increase in average collagen fibril diameter was evident from 3 to 12 week old in the WT tail (202 ± 60 nm to 243 ± 81 nm; P < 0.0001) and Achilles (160 ± 44 nm to 210 ± 63 nm; P < 0.001) tendons, accompanied by an increase in collagen fibril diameter distribution in both tendons (Figs. 6C, 6D). Mean Achilles collagen fibril diameter also increased from 3–12 week old in Hspg2^{Δ3−∕Δ3−} mice (139  ± 37 nm to 150 ± 34 nm p < 0.001) to a lesser extent than in WT mice (8 versus 31%), and significantly decreased in Hspg2^{Δ3−∕Δ3−} tail tendons (193 ± 46 nm to 84 ± 28 nm; P < 0.001). Thus differences between genotypes in mean collagen fibril diameter and distribution in tail and Achilles tendons were more marked by 12 weeks of age.