Regional molecular and cellular differences in the female rabbit Achilles tendon complex: potential implications for understanding responses to loading

Abstract

The aim of this study was: (i) to analyze the morphology and expression of extracellular matrix genes in six different regions of the Achilles tendon complex of intact normal rabbits; and (ii) to assess the effect of ovariohysterectomy (OVH) on the regional expression of these genes. Female New Zealand White rabbits were separated into two groups: (i) intact normal rabbits (n = 4); and (ii) OVH rabbits (n = 8). For each rabbit, the Achilles tendon complex was dissected into six regions: distal gastrocnemius (DG); distal flexor digitorum superficialis; proximal lateral gastrocnemius (PLG); proximal medial gastrocnemius; proximal flexor digitorum superficialis; and paratenon. For each of the regions, hematoxylin and eosin staining was performed for histological evaluation of intact normal rabbit tissues and mRNA levels for proteoglycans, collagens and genes associated with collagen regulation were assessed by real-time reverse transcription-quantitative polymerase chain reaction for both the intact normal and OVH rabbit tissues. The distal regions displayed a more fibrocartilaginous phenotype. For intact normal rabbits, aggrecan mRNA expression was higher in the distal regions of the Achilles tendon complex compared with the proximal regions. Collagen Type I and matrix metalloproteinase-2 expression levels were increased in the PLG compared to the DG in the intact normal rabbit tissues. The tendons from OVH rabbits had lower gene expressions for the proteoglycans aggrecan, biglycan, decorin and versican compared with the intact normal rabbits, although the regional differences of increased aggrecan expression in distal regions compared with proximal regions persisted. The tensile and compressive forces experienced in the examined regions may be related to the regional differences found in gene expression. The lower mRNA expression of the genes examined in the OVH group confirms a potential effect of systemic estrogen on tendon.

Introduction

Tendons are load-bearing structures that transmit forces from muscle to bone and their structure is related to their function, which differs throughout the human body. The Achilles tendon withstands large loads, while tendons in the hands exert fine movements (Kannus, 2000). Tendons or regions within a tendon are subject to compression, tension, friction or a combination thereof.

Tendons primarily consist of Collagen Type I fibers, tenocytes and proteoglycans (glycosaminoglycan chains with protein cores; Sharma & Maffulli, 2005; Scott et al. 2007; Riley, 2008). Proteoglycans are involved in Collagen Type I fibrillogenesis, and contribute to the structure and mechanical properties of tendon (Yoon & Halper, 2005; Parkinson et al. 2011). Two types of proteoglycans are present in tendons: (i) small leucine-rich proteoglycans like decorin and biglycan; and (ii) large aggregating proteoglycans like aggrecan and versican (Rees et al. 2000; Corps et al. 2006; Parkinson et al. 2011). Increased levels of aggrecan and biglycan were found in compressed regions of tendon; for example, where tendon wraps around bony prominences (Vogel et al. 1993; Benjamin & Ralphs, 1998; Parkinson et al. 2011). Increased glycosaminoglycan content in compressed regions compared with tensile regions of tendons have been reported in human tibialis posterior tendons (Vogel et al. 1993) and human biceps tendons (Berenson et al. 1996). Fibrocartilage metaplasia, a structural change, is often seen after tendon injury; therefore, it is important to understand the baseline phenotypic variations in tendon tissue.

Men and women experience tendon conditions/injuries with different frequency (Cook et al. 1998; Taunton et al. 2002; Cook & Khan, 2007). Also, men and women have differences in tendon collagen properties. In human patellar tendon, collagen (proline) fractional synthesis rate was lower for women than men (Magnusson et al. 2007), and collagen (hydroxyproline) content tended to be lower for women than men when normalized to wet weight but not to dry mass (LeMoine et al. 2009). The ultimate stress of isolated collagen fibrils from the patellar tendons of women was lower than that of men (Magnusson et al. 2007). In women, Achilles tendon ruptures demonstrated a steady increase after age 60 years, with a peak incidence at age 80 years or older, and the authors commented that the rate of rupture starts to increase after menopause (Maffulli et al. 1999). Hormones may have an influence on tendon development, injury and tendon healing. Estrogen receptors or their transcripts have been detected in human skeletal muscle (Wiik et al. 2009) and anterior cruciate ligament (Liu et al. 1996), and rabbit flexor digitorum longus, extensor digitorum, patellar and Achilles tendons (Hart et al. 1998, 2005). In rabbit tendons, the impact of pregnancy or ovariohysterectomy (OVH) was different for different tendons when evaluated using mRNA levels for genes including collagens, proteoglycans, proteinases and inflammatory mediators (Hart et al. 1998, 2005). As tendons are heterogeneous structures, it is of significance to explore the fundamental variation between intact normal and OVH tendon tissue.

The Achilles tendon complex in rabbits consists of tendons from multiple muscles. The tendon from the flexor digitorum superficialis muscle is included into the Achilles tendon complex of the rabbit (Doherty et al. 2006; Fig. 1), and it shares the paratenon – a loose areolar connective tissue that allows the tendon to move with minimal friction (Popesko et al. 1992; Jozsa & Kannus, 1997; Doherty et al. 2006) – with the medial and lateral gastrocnemius tendons. The Achilles tendon complex rotates on its way from the muscle origin to the insertion at the calcaneus (Doherty et al. 2006). The tendon of the soleus muscle in the rabbit has been shown to have a negligible amount of fibers in the Achilles tendon complex (Doherty et al. 2006). It has not been described whether these different parts of the Achilles tendon complex have varying mechanical properties or expression patterns.

Figure 1.

Posterior view of the rabbit calf muscles and tendons. The dashed line represents the dissection location where the distal part was separated from the proximal part. Inset corresponding to areas marked (A) and (B): the six regions of the Achilles tendon complex: DFDS, distal flexor digitorum superficialis; DG, distal gastrocnemius (fused lateral and medial); P, paratenon; PFDS, proximal flexor digitorum superficialis; PLG, proximal lateral gastrocnemius; PMG, proximal medial gastrocnemius. Original art by Gustav Andersson.

In this study, the expressions of genes in different anatomical regions within the rabbit Achilles tendon complex were compared. We hypothesized that compressed (distal) tendon regions would demonstrate a higher level of expression of proteoglycans compared with the tensile (proximal) regions. The paratenon is likely to show a unique expression phenotype for certain genes due to the presence of vascularization (Sharma & Maffulli, 2006) and the increased number of cells compared with the tendon proper (Archambault et al. 2001).

Differences in gene expression between intact normal and OVH rabbits were also studied in order to examine the potential effect of estrogen on gene expression in the Achilles tendon complex. We hypothesized that differences in gene expression between intact normal and OVH rabbits will be present. Collagens and genes associated with collagen regulation such as matrix metalloproteinases (MMPs) and tissue inhibitors of metalloproteinases (TIMPs) may exhibit differential expression between tendons from intact normal and OVH rabbit groups.

Materials and methods

Experiments were conducted in accordance with animal care committee approvals from University of Calgary and University of British Columbia. Achilles tendon complexes from 12 female New Zealand White rabbits (5.6 ± 0.7 kg; Riemens, St Agatha, ON, Canada) that were a minimum of 48 weeks old were used for the analysis of regional gene expression patterns and morphological characterization. Four rabbits were intact normal rabbits and eight rabbits had undergone OVH surgery at 15 weeks old. All animals were killed using an overdose of pentobarbital (Euthanyl; MTC Pharmaceuticals, Cambridge, ON, Canada).

Dissection and definition of regions

Six regions of the Achilles tendon complex were defined: distal gastrocnemius (DG); distal flexor digitorum superficialis (DFDS); proximal lateral gastrocnemius (PLG); proximal medial gastrocnemius (PMG); proximal flexor digitorum superficialis (PFDS); and paratenon. The anatomy of the rabbit Achilles complex is illustrated in Fig. 1. The Achilles tendon complex of the hindlimb was dissected from the calcaneal insertion to 3 cm proximal. The paratenon was collected separately followed by splitting the Achilles tendon complex into two parts. The distal part consisted of two tendons: the fused gastrocnemius tendons or DG; and the DFDS tendon. The proximal part consisted of three tendons: the PLG; the PMG; and the PFDS (Fig. 1). The samples for molecular analysis were snap-frozen using liquid nitrogen and stored at −80 °C until further analysis. The samples for histological analysis were fixed in 10% formalin in phosphate-buffered saline.

RNA extraction and reverse transcription-quantitative polymerase chain reaction (RT-qPCR)

RNA was extracted using the TRIspin method; this process has been previously described in detail (Reno et al. 1997). Frozen tendon tissues were pulverized using a Braun Mikro-Dismembrator (B. Braun Biotech, Allentown, PA, USA). To thaw the powdered tendon tissue, 1 mL of TRIzol Reagent (Life Techonologies, Gaithersburg, MD, USA) was added to each sample. When thawed, the samples were transferred to 1.5-mL Eppendorf tubes. Chloroform, 0.2 mL, was added followed by vortexing and centrifuging to separate the aqueous and organic phases. The RNA in the aqueous phase was purified using the RNeasy Total RNA isolation kit (Qiagen, Chatsworth, CA, USA). Quantification of total RNA was completed using SYBR Green II (FMC BioProducts, Rockland, ME, USA) fluorescent dye and comparing experimental values with standard calf liver ribosomal RNA concentrations on a Turner 450 fluorescence spectrofluorometer (Barnstead/Thermolyne, Dubuque, IA, USA). Total RNA (1 μg) was converted to cDNA using Qiagen Omniscript RT kit (Qiagen Sciences) enzyme reverse transcriptase. Real-time RT-qPCR was performed in duplicate on 19 genes (Table 1) using a BioRad iCycler (Bio-Rad, Hercules, CA, USA). The primers used (Table 1) were designed for the target rabbit genes. Values for genes were normalized to corresponding 18S values.

Table 1.

RT-PCR primers.

Target	Forward primer	Reverse primer	Base pairs
18S	TGG TCG CTC GCT CCT CTC C	CGC CTG CTG CCT TCC TTG G	360
Aggrecan	GAG GAG ATG GAG GGT GAG GTC TTT	CTT CGC CTG TGT AGC AGA TG	313
Biglycan	GAT GGC CTG AAG CTC AA	GGT TGT TGA AGA GGC TG	406
Collagen Type I	GAT GCG TTC CAG TTC GAG TA	GGT CTT CCG GTG GTC TTG TA	312
Collagen Type III	TTA TAA ACC AAC CTC TTC CT	TAT TAT AGC ACC ATT GAG AC	255
Collagen Type V	GAG GAG AAC CAG GAA TAA CC	GCA CCT TTC TCT CCG ATG CC	215
COX-2	CAA ACT GCT CCT GAA ACC CAC TC	GCT ATT GAC GAT GTT CCA GAC TCC	82
Decorin	TGT GGA CAA TGG TTC TCT GG	CCA CAT TGC AGT TAG GTT CC	419
IL-6	CCT GCC TGC TGA GAA TCA CTT	CGA GAT ACA TCC GGA ACT CCA T	51
IL-8	CAA CCT TCC TGC TGT CTC TG	GGT CCA CTC TCA ATC ACT CT	145
MMP-2	CTT CCC CCG CAA GCC CAA GTG GG	GGT GAA CAG GGC TTC ATG GGG GC	510
MMP-3	GCC AAG AGA TGC TGT TGA TG	AGG TCT GTG AAG GCG TTG TA	363
MMP-13	TTC GGC TTA GAG GTG ACA GG	ACT CTT GCC GGT GTA GGT GT	527
PRG-4	GAA CGT GCT ATA GGA CCT TC	CAG ACT TTG GAT AAG GTC TGC C	287
Tenascin C	CTG GAA TGG AGG AAT GGC	TTG GTT GTG GCT TGT TGG	116
TGF-β	CGG CAG CTG TAC ATT GAC TT	AGC GCA CGA TCA TGT TGG AC	271
TIMP-2	GTA GTG ATC AGG GCC AAA G	TTC TCT GTG ACC CAG TCC AT	416
VEGF	GGA GTA CCC TGA TGA GAT CGA	CTT TGG TCT GCA TTC ACA TTT GT	211
Versican	GAT GTG TAT TGT TAT GTG GAT CA	CAT CAA ATC TGC TAT CAG GG	310

COX, cyclooxygenase; IL, interleukin; MMP, matrix metalloproteinase; PRG, proteoglycan; TGF, transforming growth factor; TIMP, tissue inhibitor of metalloproteinase; VEGF, vascular endothelial growth factor.

For morphological characterization, hematoxylin and eosin staining was performed on 5-μm sections of tissues in the six regions of the Achilles tendon complex from one intact normal rabbit.

Statistical analysis

Linear mixed-model analysis was performed using spss (SPSS, Chicago, IL, USA) to detect statistically significant differences between the six tendon regions or the two rabbit groups. For molecular analysis, tendon tissues from both hindlimbs of three intact normal rabbits were analyzed. In the OVH group, tendon tissues from one hindlimb of eight OVH rabbits were analyzed with some exclusions: DG (n = 8); DFDS (n = 8); PLG (n = 7); PMG (n = 7); PFDS (n = 7); paratenon (n = 6). All data were log-transformed before statistical analysis to obtain normal distribution. After analysis, the data were transformed back to obtain the original format: fold change. A P-value of ≤ 0.05 was considered statistically significant for both tests. Results are given as a mean difference between the compared locations with a 95% confidence interval (CI). CI adjustments were made using Sidak correction.

Results

Anatomical findings – macroscopic examination

When dissecting the rabbit hindlimbs, the medial and lateral heads of the gastrocnemius were shown to give rise to two independent tendons that shared paratenon with the tendon from the flexor digitorum superficialis muscle. The soleus muscle was clearly seen on the ventral side of the gastrocnemius muscle, showing a slender, spindle-shaped muscle that arose from the lateral–posterior aspect of the tibia. It inserted into the Achilles tendon complex, but very few (if any) fibers extended down towards the calcaneus. At the proximal part of the Achilles tendon complex, the three individual tendons were macroscopically discernable (lateral gastrocnemius, medial gastrocnemius and flexor digitorum superficialis), but at the calcaneal insertion (distal part) the gastrocnemius tendons were fused into one tendon, found ventrally of the flexor digitorum longus tendon (Fig. 1). At this distal part, the flexor digitorum superficialis did not insert into the calcaneus but continued to course around and under the calcaneus.

Analysis of tendons from intact normal rabbits

Regional differences in tendon morphology

Histology was performed on intact normal rabbit tissue. Fibrocartilage consists of collagen fibers with a random network of interfibrillar matrix that contains high amounts of proteoglycan (Robbins et al. 1997). There are various forms of fibrocartilage, suggesting that there is a spectrum between dense fibrous connective tissue and hyaline cartilage (Benjamin & Ralphs, 1998). The DG (Fig. 2E) showed the most prominent fibrocartilage structure; in between the chondrocytes, collagen bundles were positioned in multiple orientations. The DFDS (Fig. 2D) fibrocartilage appeared highly cellular and did not show a prominent fiber structure. In the PLG (Fig. 2B), tendon tissue, as well as regions of loose connective tissue, adipose tissue and fibrocartilage, was observed. The fibrocartilage in the PFDS (Fig. 2A) had a less cellular appearance, with this region also containing more adipose tissue compared with the DG (Fig. 2E). The PMG (Fig. 2C) displayed a fibrocartilage area that appeared structurally similar to the PFDS; however, it was less cellular and a clear peritenon was visible. The distal regions showed a more pronounced fibrocartilage structure compared with the proximal regions and paratenon (Fig. 2F).

Figure 2.

Hematoxylin and eosin-stained intact normal rabbit Achilles tendon complex at a 20 × magnification. Fibrocartilage of the proximal regions (A) PFDS, (B) PLG, (C) PMG, and the distal regions (D) DFDS, (E) DG. Image (F) displays the paratenon. The fibrocartilage of the proximal regions (A–C) had a less cellular appearance with the PFDS fibrocartilage containing adipose tissue, compared with the distal regions (D, E). The black arrows in (A) indicate tenocytes and the green arrows in (E) indicate chondrocytes. The asterisks in (A) indicate adipose tissue.

Regional differences in gene expression

DG

The DG had increased aggrecan expression compared with all of the proximal regions (PLG, PMG and PFDS) in tendons from intact normal rabbits (P < 0.01; Table 2). The expression levels of Collagen Type I, MMP-2, interleukin (IL)-6 and IL-8 were significantly lower in the DG compared with the PLG and PFDS (P < 0.05; Table 2). Additionally, IL-6 expression was lower in the DG compared with the PMG (P < 0.01; Table 2). Tenascin C expression was lower in the DG compared with the PMG and PFDS (P < 0.01; Table 2).

Table 2.

Gene expression differences in regions of tendon of intact normal and OVH rabbits.

Gene	Intact normal rabbits		OVH rabbits
	Location	Mean difference (CI)	Location	Mean difference (CI)
Aggrecan	DG vs.		DG vs.
	PFDS	12.4 (2.6–57.8)*
	PLG	29.7 (6.3–138.6)*
	PMG	55.3 (11.8–259.2)*	PMG	11.9 (1.9–73.6)*
	Paratenon	609.5 (130.3–2851.8)*	Paratenon	36.0 (5.4–241.0)*
	DFDS vs.		DFDS vs.
	PFDS	12.9 (2.8–60.3)*
	PLG	30.9 (6.6–144.6)*	PLG	7.5 (1.2–45.9)†
	PMG	57.8 (12.4–270.3)*	PMG	20.2 (3.3–125.0)*
	Paratenon	635.3 (135.9–2974.7)*	Paratenon	61.0 (9.1–409.1)*
	PFDS vs.		PFDS vs.
	Paratenon	49.3 (10.6–230.9)*	Paratenon	16.9 (2.4–119.1)*
	PLG vs.		PLG vs.
	Paratenon	20.6 (4.4–96.3)*	Paratenon	8.2 (1.2–57.4)†
	PMG vs.
	Paratenon	11.0 (2.4–51.5)*
Biglycan	DFDS vs.		Paratenon vs.
	PMG	3.7 (1.0–12.9)†	DG	3.3 (1.1–10.9)†
	Paratenon	4.1 (1.2–14.3)†
IL-8	PFDS vs.
	DG	5.3 (1.3–22.3)†
	PLG vs.
	DG	5.2 (1.2–22.1)†
	Paratenon vs.		Paratenon vs.
	DFDS	5.5 (1.3–23.3)†	DFDS	4.3 (1.6–11.2)*
	DG	15.0 (3.6–63.3)*	DG	4.0 (1.5–10.5)*
	PMG	5.2 (1.2–21.7)†	PMG	8.3 (3.1–22.1)*
			PFDS	5.4 (2.0–14.3)*
			PLG	4.4 (1.6–11.9)*
MMP-2	PFDS vs.
	DG	5.2 (1.5–8.6)*
	PLG vs.
	DG	3.9 (1.1–13.9)†
	Paratenon vs.		Paratenon vs.
	DFDS	3.8 (1.1–13.5)†
	DG	10.5 (3.0–37.1)*
	PMG	6.4 (1.8–22.7)†	PMG	4.8 (1.3–18.6)†
TGF-β	Paratenon vs.		Paratenon vs.
	PMG	4.7 (1.0–21.3)†	PMG	2.4 (1.0–5.5)†
Collagen Type I	PFDS vs.		No statistical differences
	DG	10.2 (2.2–47.5)*
	PLG vs.
	DG	5.7 (1.2–26.3)†
IL-6	PFDS vs.		No statistical differences
	DG	9.1 (2.0–42.1)*
	PLG vs.
	DG	6.2 (1.3–28.5)†
	PMG vs.
	DG	4.9 (1.1–22.8)†
	Paratenon vs.
	DG	8.7 (1.9–40.4)*
Tenascin-C	PFDS vs.		No statistical differences
	DG	7.0 (1.5–32.0)*
	PMG vs.
	DG	6.5 (1.0–21.3)*
Collagen Type III	No statistical differences		DFDS vs.
			PMG	4.8 (1.9–12.5)*
			DG vs.
			PMG	2.8 (1.1–7.3)†
			PFDS vs.
			PMG	5.0 (1.9–13.4)*
			PLG vs.
			PMG	4.3 (1.5–10.7)*
			Paratenon vs.
			DG	2.8 (1.0–7.8)†
			PMG	7.9 (2.8–22.2)*
Collagen Type V	No statistical differences		DFDS vs.
			PMG	3.7 (1.4–9.5)*
			DG vs.
			PMG	3.2 (1.2–8.2)*
			PFDS vs.
			PMG	3.1 (1.2–8.3)*
			PLG vs.
			PMG	2.8 (1.1–7.5)*
			Paratenon vs.
			PMG	4.9 (1.9–13.1)*
Decorin	No statistical differences		DFDS vs.
			PMG	4.8 (1.8–12.7)*
			PFDS vs.
			PMG	3.0 (1.1–8.1)†
			PLG vs.
			PMG	3.0 (1.1–8.1)†
MMP-3	No statistical differences		DG vs.
			PMG	7.8 (1.7–35.0)*

CI, confidence interval; DFDS, distal flexor digitorum superficialis; DG, distal gastrocnemius (fused lateral and medial); IL, interleukin; MMP, matrix metalloproteinase; OVH, ovariohysterectomy; PFDS, proximal flexor digitorum superficialis; PLG, proximal lateral gastrocnemius; PMG, proximal medial gastrocnemius; TGF, transforming growth factor.

Mean differences of the regions are displayed for both rabbit groups.

^*P < 0.01

^†P < 0.05.

DFDS

The DFDS had increased aggrecan expression compared with all of the proximal regions (PLG, PMG and PFDS) in tendons from intact normal rabbits (P < 0.01; Table 2). The expression of biglycan was greater in the DFDS compared with the PMG (P < 0.05; Table 2).

Paratenon

The paratenon had lower expression levels of aggrecan compared with the two distal regions (DG and DFDS) and the three proximal regions (PLG, PMG and PFDS) in tendons from intact normal rabbits (P < 0.01; Table 2). The expression of biglycan was lower in the paratenon than the DFDS. Expression of IL-8 and MMP-2 were greater in the paratenon compared with the DG, DFDS and PMG (P < 0.05; Table 2). Paratenon expression of IL-6 was greater than the DG. The expression of transforming growth factor (TGF)-β was greater in the paratenon than the PMG.

Analysis of tendons from OVH rabbits

Regional differences in gene expression

DG

In tendons from OVH rabbits, aggrecan expression was higher in the DG compared with the PMG (P < 0.01; Table 2). Additionally, the DG exhibited an increased expression compared with the PMG for Collagen Type III, Collagen Type V, decorin and MMP-3 (P < 0.05; Table 2).

DFDS

The DFDS had increased aggrecan expression compared with the PLG and PMG in tendons from OVH rabbits (P < 0.05; Table 2). Also, the DFDS demonstrated increased expression of Collagen Type III, Collagen Type V and decorin compared with the PMG (P < 0.01; Table 2).

PMG

The PMG had decreased expression of Collagen Type III, Collagen Type V and decorin compared with both the PFDS and PLG in tendons from OVH rabbits (P < 0.05; Table 2).

Paratenon

The paratenon had decreased expression of aggrecan compared with both distal regions (DG and DFDS) in tendons from OVH rabbits (P < 0.01; Table 2). Additionally, aggrecan expression was lower in the paratenon compared with the PLG and PFDS (P < 0.05; Table 2). The expression of biglycan was greater in the paratenon than the DG (P < 0.01; Table 2). Expression of MMP-2 and TGF-β were greater in the paratenon compared with the PMG (P < 0.05; Table 2). The expression of Collagen Type III and Collagen Type V was greater in the paratenon than the PMG. Additionally, the expression of Collagen Type V was greater in the paratenon than the DG (P < 0.01; Table 2).

Comparisons of tendons from both intact normal and OVH rabbits

To test whether the surgical menopause had an effect on gene expression values, expression levels for the genes in question were compared between the intact normal and OVH groups. Per gene, the data for the PLG, PMG and PFDS were pooled and compared with the pooled DG and DFDS data. A significantly lower expression of 10 genes of the 18 investigated was detected in the OVH rabbits (Table 3).

Table 3.

Gene expression data for tendons from intact normal and OVH rabbits of the genes that exhibit significant differences between the rabbit groups.

Gene	Mean intact normal (CI)	Mean OVH (CI)
Aggrecan	297.2 (145.9–606.7)	28.9 (15.1–55.5)*
Biglycan	9.5 (6.5–13.9)	5.3 (4.1–7.0)†
Collagen Type I	19.9 (12.1–32.8)	7.1 (5.0–10.2)*
COX-2	14.7 (6.8–32.1)	4.7 (2.9–7.7)†
Decorin	25.1 (11.5–54.8)	6.1 (3.7–10.1)†
IL-6	27.4 (11.6–64.7)	8.8 (5.1–15.2)†
TGF-β	10.2 (5.5–18.9)	4.0 (2.7–6.0)†
TIMP-2	8.8 (5.5–14.1)	4.4 (3.2–6.0)†
Versican	10.8 (6.5–17.7)	2.2 (1.6–3.0)*
VEGF	7.4 (4.6–11.9)	4.2 (3.0–5.8)†

The data are expressed in mean relative quantity with confidence interval (CI).

COX, cyclooxygenase; IL, interleukin; OVH, ovariohysterectomy; TGF, transforming growth factor; TIMP, tissue inhibitor of metalloproteinase; VEGF, vascular endothelial growth factor.

^*P < 0.01

^†P < 0.05.

A 10-fold lower expression of aggrecan in mean relative quantity was found in the OVH group. Biglycan showed almost a twofold decrease in the mRNA expression in the OVH rabbits. The other proteoglycans; decorin and versican, demonstrated four and five times less mRNA in the OVH group. The Collagen Type I expression was two and a half times lower in the OVH rabbit tissue. The enzyme cyclooxygenase-2 was downregulated approximately three times more in the OVH group. IL-6 was detected in a lower quantity in the OVH group compared with the normal group. A twofold decrease in mRNA expression of TIMP-2 was observed in the OVH tendon tissue. TGF-β and vascular endothelial growth factor mRNA were also detected in lower amounts in the OVH vs. intact normal group.

Despite these lower levels of expression, consistent regional variations were observed between the two rabbit groups (Table 2). Aggrecan expression was greater in the DG compared with the PMG, and in the DFDS compared with the PLG and PMG for both intact normal and OVH rabbits (Table 2). In both rabbit groups, aggrecan expression was lower in the paratenon than in the DG, PLG, DFDS and PFDS. The expression of MMP-2, TGF-β and IL-8 was greater in the paratenon than the PMG. In addition, the expression of IL-8 was greater in the paratenon than in both distal regions (DG and DFDS).

No statistically significant differences were detected in gene expression values between the tendons from the left and right hindlimb of the rabbits in either the intact normal or OVH rabbits. Therefore, the specific hindlimb used was not considered a factor of importance.

Discussion

Differences were found in gene expression in the various regions of the Achilles tendon complex in both rabbit groups, which may be related to the loads experienced in the various regions, in other words, their tensile and compressive environments. The compressed (distal) regions of the intact normal Achilles tendon complex displayed a more fibrocartilaginous phenotype, as evidenced by higher aggrecan and biglycan expression levels as well as fibrocartilaginous tissue morphology compared with the tensile (proximal) regions. The mRNA levels of the tendons from the OVH rabbits displayed reduced expression of 10 of 18 genes examined, which suggests that systemic estrogen may have a general effect on tenocyte metabolism.

Influence of tendon region on gene expression

The areas of tendons that wrap around bony prominences, thus experiencing compression, have been shown to exhibit increased levels of aggrecan and biglycan (Vogel et al. 1993; Benjamin & Ralphs, 1998; Parkinson et al. 2011). Fibrocartilaginous morphology is associated with compressed regions (Benjamin & Ralphs, 1998; Parkinson et al. 2011), and this is also illustrated in Fig. 2 where a more cartilage-like structure is displayed in the compressed (distal) regions. In the present study, in both rabbit groups, aggrecan exhibited increased expression levels in the distal regions compared with most of the proximal regions. This increase can likely be attributed to its location in the Achilles tendon complex as both distal regions are closely situated to the bone and partly compressed by it during loading.

The findings presented are consistent with previous findings examining the calcaneal tendon of male rats (Marqueti et al. 2012). The expression levels of Collagen Type I and MMP-2 were increased in the proximal region compared with the distal region of the calcaneal tendon for male rats in their sedentary group (Marqueti et al. 2012). In the current study, the expression levels of Collagen Type I and MMP-2 were increased in the PLG compared with the DG for female rabbits in the intact normal group. In addition to regional differences in gene expression, the rat calcaneal tendon model has been used to investigate regional differences in total MMP-2 activity. In the rat calcaneal tendon, total MMP-2 activity was not different comparing the proximal with the distal region for male rats (Marqueti et al. 2008), lower in the proximal compared with the distal region for female rats (Pereira et al. 2010), and higher in the proximal compared with the distal region for ovariectomized female rats (Pereira et al. 2010). These activity findings emphasize the importance of gender/hormonal effects on tendon regional differences.

Collagen Type III and V expression levels were lower in the PMG compared with most other regions in the OVH rabbits. Similarly, the PMG displayed a decrease in decorin expression compared with the PLG, PFDS and DFDS in the OVH rabbits. Remarkable is that of the three proximal tendon regions, the PMG is the only region exhibiting a lower expression of Collagen Type III, Collagen Type V and decorin compared with most other regions. This could be due to the location of the PMG tendon; it initiates between the PLG and PFDS and rotates to the side of the Achilles tendon complex. Decorin is the most abundant proteoglycan in tensile regions of tendon (Samiric et al. 2004; Parkinson et al. 2011), and the PMG may not be experiencing the same amount of tensile force as the PLG and PFDS.

OVH effects on gene expression

The effect of estrogen on proteoglycan levels in tendon, to our knowledge, has not been researched extensively. Among the genes investigated in this study, the aggrecan mRNA expression showed the biggest difference between the two groups, even though regional variations were preserved. Aggrecan had a 10-fold lower expression in the tendons from OVH rabbits compared with the intact normal rabbits. Interestingly, all of the other proteoglycans investigated (biglycan, decorin and versican) also had decreased expression in the tendons from OVH rabbits compared with those from intact normal rabbits.

In a longitudinal study, Maffulli et al. (1999) commented that the rate of Achilles tendon rupture starts to increase after menopause. This likely suggests that systemic estrogen may play a protective role for tendon injuries. Comparing patellar tendons from postmenopausal women using hormone replacement therapy (HRT) with those from non-users of HRT, collagen (proline) fractional synthesis rate was increased but Collagen Type I (pro-collagen type I N-terminal propeptide: PINP) synthesis was not (Hansen et al. 2009). This may indicate differences in the synthesis rates of soluble and insoluble collagen or Collagen Type I and Collagen Type III (Hansen et al. 2009; Pingel et al. 2012b). In our study, the Collagen Type I mRNA level in the Achilles tendon complex of the OVH rabbits was significantly reduced compared with intact normal tendon levels.

Interleukin-6 production is repressed by estrogen due to binding of activated estrogen to transcription factors that prevent binding to DNA, as demonstrated in human osteoblasts and bone marrow stromal cells (Stein & Yang, 1995). In human Achilles tendon peritendinous tissue, Collagen Type I (PINP) synthesis was increased with infusion of IL-6 (Andersen et al. 2011). Serum IL-6 concentrations were significantly lower in postmenopausal women using HRT compared with non-users of HRT (Scheidt-Nave et al. 2001). However, Maret et al. (1999) showed that estrogen had no effect on IL-6 production in rat smooth muscle cells. The role of IL-6 may be species-and tissue/cell-specific, as the current study found three times lower mRNA expression of IL-6 in the Achilles tendon complex from OVH rabbits compared with intact normal rabbits.

This study focused on the gene expression differences in the different regions of the rabbit Achilles tendon complex. It remains to be seen whether such regional differences are observed when analyzing protein levels or enzyme activity. In the rat calcaneal tendon model, regional differences in total MMP-2 activity were observed (Marqueti et al. 2008; Pereira et al. 2010). In human patellar tendon collagen fascicles, regional (anterior vs. posterior) differences in mechanical properties were observed (Haraldsson et al. 2005). A larger future study could incorporate gene expression, enzyme activity, protein level and mechanical property assessments to increase our understanding of these regional differences and their relationship to mechanical loading environments.

In agreement with the hypotheses, changes in gene expression between tendon regions and between rabbit groups were demonstrated even though small sample size was a limitation of this study. The compressed regions showed higher mRNA expression levels of proteoglycans compared with the tensile regions. Expression of Collagen Type I, which plays a role in tensile load-bearing, was greater in the tensile PLG region compared with the compressed DG region. These findings suggest that the mRNA expression levels differ in the tendon regions, based on the type of load experienced in those regions. All proteoglycans examined in this study displayed a lower mRNA expression in the OVH group, despite preserving the regional differences, illustrating that estrogen likely plays a role in the expression of genes that are important for tendon homeostasis.

In summary, the different regions of the Achilles tendon complex in rabbits have different gene expressions related to the type of load that is exerted upon them. This should be taken into consideration whenever the tendon is used in research as biopsies may differ depending on where in the tendon complex it is taken. One should be careful in applying the findings of these regional differences in the rabbit to the human Achilles tendon as the basic anatomy differs, most notably the lack of a flexor digitorum superficialis tendon and the more pronounced supply of tendon fibers from the soleus muscle, in the human Achilles tendon complex. In order to delineate regional differences in the human Achilles tendon, further studies are required. A recent study where human Achilles tendons were biopsied from a more proximal region within the same tendon for healthy tissue to be compared with tendinopathic tissue from a more distal region of the same tendon (Pingel et al. 2012a) suggests that such regional difference studies could be of clinical interest.

Conflict of interest

No conflict of interest.