Histochemistry and Cell Biology

, 136:437

Colocalization in vivo and association in vitro of perlecan and elastin

Authors

  • Anthony J. Hayes
    • BioImaging Unit, Cardiff School of BiosciencesUniversity of Cardiff
  • Megan S. Lord
    • Graduate School of Biomedical EngineeringUniversity of New South Wales
  • Susan M. Smith
    • Raymond Purves Bone and Joint Research Laboratories, Institute of Bone and Joint Research, Kolling Institute of Medical ResearchUniversity of Sydney, The Royal North Shore Hospital
  • Margaret M. Smith
    • Raymond Purves Bone and Joint Research Laboratories, Institute of Bone and Joint Research, Kolling Institute of Medical ResearchUniversity of Sydney, The Royal North Shore Hospital
  • John M. Whitelock
    • Graduate School of Biomedical EngineeringUniversity of New South Wales
  • Anthony S. Weiss
    • School of Molecular BioscienceUniversity of Sydney
    • Raymond Purves Bone and Joint Research Laboratories, Institute of Bone and Joint Research, Kolling Institute of Medical ResearchUniversity of Sydney, The Royal North Shore Hospital
Original Paper

DOI: 10.1007/s00418-011-0854-7

Cite this article as:
Hayes, A.J., Lord, M.S., Smith, S.M. et al. Histochem Cell Biol (2011) 136: 437. doi:10.1007/s00418-011-0854-7

Abstract

We have colocalized elastin and fibrillin-1 with perlecan in extracellular matrix of tensional and weight-bearing connective tissues. Elastin and fibrillin-1 were identified as prominent components of paraspinal blood vessels, and posterior longitudinal ligament in the human fetal spine and outer annulus fibrosus of the fetal intervertebral disc. We also colocalized perlecan with a synovial elastic basal lamina, where the attached synovial cells were observed to produce perlecan. Elastin, fibrillin-1 and perlecan were co-localized in the intima and media of small blood vessels in the synovium and in human fetal paraspinal blood vessels. Elastic fibers were observed at the insertion point of the anterior cruciate ligament to bone in the ovine stifle joint where they colocalized with perlecan. Elastin has not previously been reported to be spatially associated with perlecan in these tissues. Interactions between the tropoelastin and perlecan heparan sulfate chains were demonstrated using quartz crystal microbalance with dissipation solid phase binding studies. Electrostatic interactions through the heparan sulfate chains of perlecan and core protein mediated the interactions with tropoelastin, and were both important in the coacervation of tropoelastin and deposition of elastin onto perlecan immobilized on the chip surface. This may help us to understand the interactions which are expected to occur in vivo between the tropoelastin and perlecan to facilitate the deposition of elastin and formation of elastic microfibrils in situ and would be consistent with the observed distributions of these components in a number of connective tissues.

Keywords

PerlecanElastinHeparan sulfateHSFibrillin-1

Introduction

Perlecan is a modular proteoglycan predominantly substituted with heparan sulfate (HS) with well-documented roles in extracellular matrix organization and stabilization, manifested through its ability to interact with a diverse repertoire of matrix components (Iozzo 1994, 1998, 2005; Iozzo et al. 1994; Melrose et al. 2008a). Perlecan can bind a subset of growth factors and morphogen thereby protecting them from proteolytic attack in free solution, extending their biological half-life and strategically sequestering them in the pericellular matrix around connective tissue cells or in morphogen or growth factor gradients within developmental tissues (Whitelock et al. 2008). Perlecan participates in angiogenesis and vasculogenesis (Bix and Iozzo 2008; Costell et al. 2002; Iozzo and San Antonio 2001; Melrose et al. 2008a) and is a key component of basement membranes (Iozzo 2005; Murdoch and Iozzo 1993; Murdoch et al. 1994). Elastin and fibrillin-1 are assembled into fine microfibrils, intermediate and mature fibers, sheets or laminae. Elastic fibers have been immunolocalized in flexor (Ritty et al. 2002) and tail tendon (Caldini et al. 1990) and despite the relatively inextensible nature of tendons and ligaments the elastic fibers are considered to convey important compliance properties to these tissues (de Carvalho et al. 1994; Kannus 2000; Montes 1996). In the present study we have colocalized perlecan and elastin and perlecan and fibrillin-1 in the lumenal and medial tissues of synovial and paraspinal blood vessels, in the anterior cruciate ligament (ACL) insertion site to bone, and posterior longitudinal spinal ligament. Perlecan and elastin were also colocalised in an elastic basal lamina in synovium. We have also shown that elastin is an important component of the developing outer annulus fibrosus in the human fetal and newborn ovine intervertebral disc.

Elastic fibers are macromolecular extracellular matrix assemblies that endow dynamic connective tissues such as arteries, ligaments and tendon with elastic recoil properties (Montes 1996). Tropoelastin is the soluble, secreted form of elastin. It comprises alternating hydrophobic and lysine-rich crosslinking domains and acts as a ‘block co-polymer’ in vitro, undergoing a unique form of self assembly, termed coacervation (Clarke et al. 2005; Kielty et al. 2007). The major structural components of elastic fibers are elastin and fibrillin-1 (Kielty et al. 2007; Stephan et al. 2006; Williamson et al. 2007) however several additional molecules are also associated with their assembly. Recent mass spectrometry and molecular fishing have identified up to 200 proteins, including perlecan, associated with microfibrils. Interactions between (a) tropoelastin and perlecan-HS chains and between (b) fibrillin-1 and perlecan-HS chains each may contribute to the assembly of elastic fibrillar components (Broekelmann et al. 2005; Brown-Augsburger et al. 1994; Cain et al. 2005; Gheduzzi et al. 2005; Sasaki et al. 1999) and are considered important in the assembly and stabilization of basement membranes (Henderson et al. 1996; Hirani et al. 2007; Tiedemann et al. 2001; Tiedemann et al. 2005; Zhang et al. 1994). These models are consistent with the co-localization of perlecan, fibrillin-1 and elastin we observed in the present study in the synovial elastic lamina, and intimal and medial regions of blood vessels. The present study is the first of its kind to colocalize elastin, perlecan and fibrillin-1 in such morphologically diverse tissues. This study also quantified interactions between perlecan and tropoelastin, the secreted form of elastin using quartz crystal microbalance with dissipation (QCM-D), these interactions with tropoelastin rely on perlecan’s HS chains and core protein (Clarke et al. 2006; Clarke et al. 2005; Toonkool et al. 2001; Wise and Weiss 2009).

Materials and methods

Antibodies, enzymes and reagents

Mouse anti-bovine α-elastin (MAb BA4) (Grosso and Scott 1993; Maier et al. 1994; Wrenn et al. 1986), low molecular weight heparin, chondroitinase-ABC were obtained from Sigma-Aldrich. Menzel and Glaser SuperFrost ultraPlus, positively charged microscope sides were obtained from Fisher Scientific. Biotinylated anti-mouse IgG secondary antibody and horseradish peroxidase conjugate were obtained from Dako. Histochoice was from Amresco. NovaRED substrate was obtained from Vector Laboratories. Rat monoclonal anti-perlecan domain-IV (mAb A7L6) (Couchman and Ljubimov 1989; Horiguchi et al. 1989; Ljubimov et al. 1992) was obtained from Abcam. A polyclonal antibody raised in sheep to the N-terminal proline rich region of fibrillin-1 was a kind gift from Prof Penny Handford, University of Oxford. Alexa 594 conjugated donkey anti-rat and Alexa 488 anti-mouse secondary antibodies were obtained from Molecular Probes (Invitrogen). Heparinase III (EC 4.2.2.8) was purchased from Seikagaku Corp.

Tissues

Ovine tissues were obtained from Merino sheep aged 2 days or 2 years under institutional ethics. A 14-week-old human fetal lumbar spinal segment was obtained with informed consent at termination of pregnancy with ethical approval from the Human Care and Ethics Review Board of The Royal North Shore Hospital.

Histological processing of specimens

Full thickness mid coronal tissue blocks of ovine femoral condyles from 2-year-old sheep were fixed 48 h in 10% neutral buffered formalin pH 7.2 and decalcified in 10% formic acid, dehydrated in graded ethanol solutions and embedded in Paraplast wax blocks. 4 or 7 μm microtome sections were cut and attached to SuperFrost Plus glass microscope slides (Menzel-Glaser, Germany), deparaffinized in xylene (2 changes×5 min), and rehydrated through graded ethanol washes (100–70% v/v) to water (Melrose et al. 2004). Samples of synovium from the suprapatellar fat pad of 2-year-old sheep were fixed for 24 h in 10% v/v neutral buffered formalin and processed into paraffin blocks. A 14-week-old human fetal thoracolumbar spinal segment was fixed for 24 h in Histochoice and single embedded in paraffin (Melrose et al. 2008b) and 7 μm vertical longitudinal parasaggital sections were made.

Bright-field immunolocalization of aggrecan mid coronal sections of ovine stifle joints

Endogenous peroxidase activity was blocked by incubating sections with 3% H2O2 for 5 min. After washing in water non-specific binding sites were blocked with 10% swine serum for 10 min. Sections were predigested with chondroitinase ABC (0.25 U/ml) for 1 h at 37°C in 0.1 M Tris–HCl, 0.03 M sodium acetate buffer (pH 8.0), then mAb 2B6 (1:5,000 dilution), which recognises the 4-sulfated (C-4-S) linkage region of chondroitin sulfate (CS) chains after chondroitinase ABC digestion (Couchman et al. 1984) overnight at 4°C. After washing in TBS-Tween the primary Ab was localized using biotinylated anti-mouse IgG, and then horseradish peroxidase conjugated streptavidin was used to visualize the tissue immune complexes with Nova RED substrate for color development (Melrose et al. 2006; Melrose et al. 2005a, 2005b). Control sections were prepared in which the authentic primary antibody was either omitted or replaced with an irrelevant isotype-matched mouse IgG directed against Aspergillus niger glucose oxidase.

Laser scanning confocal microscopy: dual localisation of perlecan and elastin and fibrillin-1 and perlecan

Tissue sections (7 μm) were predigested with 0.4 U/ml chondroitinase ABC in 100 mM Tris acetate buffer (pH 6.5) for 1 h at 37°C followed by digestion in proteinase-K (500 μg/ml) for 15 min at room temperature then washed in PBS containing 0.001% Tween 20 pH 7.4 (PBS-Tween) and blocked with normal donkey serum (1:20 dilution) for 30 min at room temperature. An anti-perlecan rat mAb (A7L6, 1:50 dilution) was then incubated with the sections overnight at 4°C. Sections were washed and an Alexa-594 conjugated donkey anti-rat secondary antibody (1:200 dilution) was applied for 1 h at room temperature. After washing, sections were blocked with normal goat serum (1:20 dilution) for 30 min at room temperature before incubation with mouse anti-elastin mAb BA4 (1:20 dilution) or anti-fibrillin-1 (proline rich region) pAb (1:20 dilution). Sections were washed before application of Alexa 488-conjugated goat anti-mouse (1:200 dilution) or donkey anti-goat (1:200 dilution) secondary antibodies as appropriate for 1 h at room temperature. Sections were mounted over coverslips in Vectashield mountant containing DAPI. Tissue sections were scanned with a Leica TCS SP2 AOBS laser scanning confocal microscope using 40× and 63× oil immersion objectives. Samples were scanned by sequential recordings of DAPI (Ex max: 359 nm; Em. max: 461 nm), Alexa 488 (Ex max: 488 nm; Em max: 520 nm) or Alexa 594 (Ex max 594 nm; Em max: 618 nm). Z-stacks of optical sections were taken through the full thickness of tissue sections at 0.4–0.6 μm increments and maximum intensity type reconstructions prepared from image stacks using Leica Confocal Software. Areas of colocalization of perlecan and elastin, or perlecan and fibrillin-1 labeled with red (Alexa 594) or green (Alexa 488) fluorochromes, respectively, were evident in merged confocal images as yellow areas.

Quartz crystal microbalance assays

The binding of perlecan to tropoelastin was monitored and quantified in a QCM-D monitoring axial flow chamber (Q-Sense AB, Sweden) internally maintained at 20.0 ± 0.1°C. Gold QCM-D crystals were mounted in the axial flow chamber then frequency (f) and dissipation (D) measurements versus time were recorded. Dulbecco’s phosphate buffered saline, pH 7.2 (DPBS) was injected into the chamber and allowed to stabilize before being replaced with 20 μg/ml recombinant human tropoelastin (Martin et al. 1995; Wu et al. 1999) which was allowed to adsorb for 30 min. The chamber was then rinsed with DPBS for f and D stabilization. DPBS was replaced with 0.1% w/v casein in DPBS to block free gold sensor crystal binding sites and allowed to stabilize for 30 min before rinsing with DPBS and stable f and D traces were re-established. Immunopurified human coronary artery endothelial cell (HCAEC)-derived perlecan (10 μg/ml) (Knox et al. 2001; Whitelock et al. 1999) or low molecular weight heparin (Sigma Aldrich) was injected and allowed to interact for 30 min. Crystals were then rinsed with DPBS followed by 1 M NaCl with stable f and D traces established before subsequent buffer changes to assess the reversibility of the binding. The same approach was repeated by coating the gold QCM-D sensor crystals with 10 μg/ml HCAEC perlecan or HCAEC perlecan that had been pretreated with 0.01 U/ml heparinase III (EC 4.2.2.8, Seikagaku Corp.) for 16 h at 37°C. The perlecan coated crystals were then exposed to either tropoelastin or a truncated form of tropoelastin (exons 2-25), ELN27-540 (Wu et al. 1999). Adsorbed mass estimates were derived by the Voigt model from three separate experiments (Voinova et al. 1999).

Results

Immunolocalization of perlecan, fibrillin-1 and elastin in connective tissues

Elastin was prominent on the luminal surfaces of paraspinal blood vessels of a 14-week -gestational age spinal segment of human fetal spine (Fig. 1a–c). Elastin and fibrillin-1 were prominent components of the posterior longitudinal ligaments; small elastin and fibrillin-1 fibers were also seen in the outer annulus fibrosus of the intervertebral disc (Fig. 2a–e). Perlecan was also immunolocalized to the pericellular matrix of fibrochondrocytes in the outer AF (Fig. 2b), fibrillin-1 fibrils were also found attached to cells in the posterior longitudinal ligament and outer annulus fibrosus where it may have regulatory roles to play in mechanotransductive processes in extracellular matrix development and homeostasis (Fig. 2 b–e). Laser scanning confocal fluorescence microscopy showed that elastin (Fig. 3d, h) and fibrillin-1 (Fig. 4 c, g) colocalized with perlecan in the paraspinal blood vessels. Paraspinal blood vessels occurred in two differing morphologies with some vessels displaying distinct elastin and fibrillin-1 intimal colocalization with perlecan while in others diffuse colocalization in the adventitial tissues was also evident (Figs. 3, 4). Higher power examination of a selected paraspinal blood vessel clearly showed the distinct intimal colocalization of perlecan and elastin and a more diffuse medial and adventitial localization pattern (Fig. 5d). Elastin also colocalized with perlecan in the elastic basal lamina upon which synovial cells were attached (Fig. 6e) and which displayed prominent localization of perlecan, and in blood vessels of the synovium (Fig. 7e). Vertical coronal sections of the femoral condyles containing the anterior cruciate ligament attachment regions of the ovine stifle joints revealed short elastic fibers that colocalized with perlecan (Fig. 8e). Fibrillin-1 was also immunolocalized when associated with the intimal and adventitial components of small blood vessels in the ACL attachment regions to the femoral condyle (Fig. 9c–f). Assessment of the anti-mouse and anti-rat secondary detection reagents used for the confocal localizations clearly showed a lack of cross reactivity between these under the conditions employed in this study using human foetal disc, ovine synovium and intervertebral disc (Figs. 10, 11). This also demonstrated that the well-known autofluorescent properties of dense collagenous connective tissues were not a significant problem in the tissues examined in this study under the experimental conditions we employed. Autofluorescence can occur right across the visible range and poses a particular problem for conventional epifluorescence imaging using mercury vapour lamp illumination and broad band pass filters that image light emitted from the full thickness of the tissue section. This is much less of a problem for confocal microscopy as only a thin (0.5 μm) confocal optical slice is imaged at a time, The fluorescent probes were very specifically excited by laser at (or near) their absorption maxima. Fluorescent emissions are selectively recorded via narrow-band detection centred on the emission maxima for each probe using tuneable spectral detectors which allows highly specific recordings to be made of the emission profiles of each fluorescent probe. The outer annular lamellae of newborn sheep showed a distribution of perlecan along the lines of force within the lamellar layers with annular fibrochondrocytes (Fig. 12a, e). Elastin was also found in the annular layers in close proximity to perlecan (Fig. 12d, h). Higher magnification suggested that elastin was deposited into the matrix by the cells through perlecan-containing, funnel-like projections (Fig. 13a–d). Elastic fibers were only occasionally found in the nucleus pulposus (data not shown).
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig1_HTML.jpg
Fig. 1

Fluorescent immunolocalization of elastin (green) in three paraspinal blood vessels (ac) in a 14 week-old-gestational age human fetal spinal specimen. Cell nuclei were stained with DAPI (blue)

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig2_HTML.jpg
Fig. 2

Fluorescent immunolocalization of elastin (a) dual localization of fibrillin-1 and perlecan (b) and fluorescent immunolocalization of fibrillin-1 (ce) in the posterior longitudinal ligament (PLL) and outer AF (OAF) of a 14 week-old-gestational age human fetal spinal specimen. Perlecan was expressed in close proximity to fibrillin-1 fibres in the outer AF (b) and fibrillin-1 fibrils attached the cells to their adjacent extracellular matrix. Cell nuclei were stained with DAPI (blue)

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig3_HTML.jpg
Fig. 3

Fluorescent colocalization of perlecan and elastin in two selected paraspinal blood vessels: (ad) vessel 1 (eh) vessel 2 of a 14 week old gestational age human foetal spinal segment. Areas of colocalization within the specimen are evident as yellow coloured areas (superimposition of red and green) in the merged image. All scale bars represent 50 μm. The boxed area in segment d is provided at higher magnification in Fig. 5

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig4_HTML.jpg
Fig. 4

Fluorescent colocalization of perlecan and fibrillin-1 in two selected paraspinal blood vessels: ad vessel 1 eh vessel 2 of a 14 week-old-gestational age human foetal spinal segment. Areas of colocalization within the specimen are evident as yellow coloured areas (superimposition of red and green) in the merged image. All scale bars represent 50 μm. The boxed area in segment d is provided at higher magnification in Fig. 5

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig5_HTML.jpg
Fig. 5

Fluorescent colocalization of perlecan (a) and elastin (b) predominantly to the adventitia of a small paraspinal blood vessel with sparse colocalization in the medial tissue. Cell nuclei are stained blue using DAPI (c). Segments ad are higher power images of the boxed area depicted in Fig. 3d

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig6_HTML.jpg
Fig. 6

Fluorescent colocalization of perlecan and elastin in a basal lamina in a 2 year old sample of ovine synovium (ae). Low power magnification H & E images are presented in a as a guide with the boxed areas presented at higher magnification in the fluorescent colocalization images in photosegments (be). Perlecan was prominently colocalized with the elastic lamina in the synovial intima in which reside synoviocytes capable of pericellular expression of perlecan (arrows, b). Elastin is prominently colocalized with perlecan in the adventitial and medial tissues of the selected synovial blood vessel (asterisk, e)

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig7_HTML.jpg
Fig. 7

Fluorescent colocalization of perlecan and elastin in a synovial blood vessel (e). A low power magnification H & E image is presented in a as a guide with the boxed area presented at higher magnification in the fluorescent colocalization images in photo segments (e). Perlecan was prominently colocalized with the elastic lamina in the synovial intima in which reside synoviocytes capable of pericellular expression of perlecan. Elastin is prominently colocalized with perlecan in the adventitial and medial tissues of the selected synovial blood vessel

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig8_HTML.jpg
Fig. 8

Immunolocalization of the unsaturated stub epitopes of the chondroitin-4-sulfate side chains of aggrecan using mAb 2-B-6 following chondroitinase ABC pre-digestion of a mid coronal section of ovine femoral condyle (a) delineating the extent of the articular and growth plate cartilages and the anterior cruciate ligament (ACL) attachment points (boxed area). The boxed area in a is shown at higher magnification in the confocal fluorescent images presented in photo segments which depict the localization of b perlecan c elastin and d cell nuclei in the ACL sections. Perlecan and elastin are colocalized to short fibrillar material in the ACL in the merged image (e)

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig9_HTML.jpg
Fig. 9

Haematoxylin and eosin stained macromolecular view (a) of a femoral condyle with the boxed area depicted at higher magnification in b delineating the extent of the articular and growth plate cartilages and the anterior cruciate ligament (ACL) attachment points (boxed area). Selected areas within the boxed area in b are shown at higher magnification in the confocal fluorescent images presented depicting the localization of fibrillin-1 in three blood vessels (cf). The boxed areae in segment c is depicted at higher magnification in segment e. Cell nuclei are stained with DAPI (blue)

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig10_HTML.gif
Fig. 10

Demonstration of the lack of cross-reactivity between the fluorescently labeled anti-mouse and anti-rat secondary detection systems used in confocal microscopy in this study using human foetal disc and ovine synovium. A lack of autofluorescence in the tissues is also evident

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig11_HTML.gif
Fig. 11

Demonstration of the specificity of the fluorescently labeled anti-mouse and anti-rat secondary detection systems used in confocal microscopy on ovine disc (a) and ovine synovium (b). A lack of significant autofluorescence in the tissues is also demonstrated

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig12_HTML.jpg
Fig. 12

Fluorescent immunolocalization of perlecan (a, e) and elastin (b, f) in the outer annulus fibrosus of two newborn intervertebral discs ovine L2L3 (ad) and L3L4 (eh). Perlecan and elastin are localized pericellularly parallel with collagen fibre bundles running along the major axis of the annular lamellae. Cell nuclei are stained blue with DAPI (c, g). The boxedareas in segments d and h are also presented at higher magnification in Fig. 9

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig13_HTML.jpg
Fig. 13

Fluorescent immunolocalization of perlecan (red, small arrows) in close proximity to areas of elastin deposition (green, long arrows) around annular fibrochondrocytes in the outer annulus fibrosus of a newborn ovine intervertebral disc (ad). These are higher power magnifications of the boxedareas indicated in Fig. 8d, h. Large arrows depict areas of elastin deposition in proximity to pericellular funnel-like projections of perlecan (small arrows). The elastin is deposited along the major axis of the annular lamellae parallel to the collagen fibre bundles which are a prominent feature in the organization of this tissue. All scale bars represent 10 μm

Interaction of perlecan and elastin in vitro: quartz crystal microbalance with dissipation monitoring

The HS side chains of perlecan (Perlecan-HS) immobilized onto a gold QCM-D crystal bound directly to tropoelastin causing a decrease in QCM-D frequency and an increase in dissipation (Fig. 14a). The interaction between the tropoelastin bound onto a gold QCM-D crystal and perlecan-HS added in solution phase resulted in the deposition of an approximately 1.5 nm layer. This layer was completely removed by the addition of 1 M NaCl indicating that reversible electrostatic interactions were responsible for its assembly. The same type of interaction was observed between the tropoelastin and heparin, although the heparin layer was thinner (1.1 nm) than for perlecan-HS (Fig. 14b). The interaction between the solution phase tropoelastin with perlecan core protein and HS side chains immobilized on the sensor crystal resulted in decrease in frequency and dissipation (Fig. 14a) indicating that tropoelastin binding to perlecan-HS resulted in a more rigid layer than perlecan-HS alone, as would be expected from a change in hydration. The tropoelastin layer bound to perlecan was approximately 1.3 nm thick and this interaction was completely reversible with the addition of 1 M NaCl, indicating the electrostatic nature of the layer. The interaction between tropoelastin and perlecan protein core devoid of HS chains resulted in a decrease in frequency and an increase in dissipation which is consistent with binding; however, the thickness of the 3 nm layer was double than that produced by tropoelastin bound to perlecan-HS (Fig. 14c). Furthermore, this interaction was stronger as it was not completely reversible with 1 M NaCl, which only resulted in the release of some of the bound tropoelastin. Interactions between tropoelastin and perlecan-HS were also investigated using a truncated form of tropoelastin, ELN27-540, which does not contain the C-terminal domain of tropoelastin. The ELN27-540 layer bound to perlecan was only 0.4 nm thick indicating that the C-terminal domain of tropoelastin was required for interactions with perlecan (Fig. 14d). These interactions between tropoelastin and perlecan indicate a role for both the protein core and the HS side chains in the aforementioned assembly processes.
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0854-7/MediaObjects/418_2011_854_Fig14_HTML.gif
Fig. 14

Characterization of interaction between the tropoelastin and HCAC perlecan determined by QCM-D at 20 ± 0.1°C. a Changes in dissipation (triangle Dissipation) versus changes in frequency (triangle Frequency) are presented for the 3rd overtone. These data were analyzed using the Voigt model to determine b layer thickness of either perlecan or heparin interacting with tropoelastin adsorbed onto gold QCM-D crystals c the layer thickness of tropoelastin interacting with either perlecan-HS or perlecan treated with heparinase III to remove it’s HS chains [Perlecan (-HS)] and d the layer thickness of ELN27-540 interacting with perlecan-HS. The data presented in the histograms is mean data of three separate experiments, the error bars are standard deviations

Discussion

In the present study, fibrillin-1 and elastin were both found to localize with perlecan in blood vessels of variable morphology, in connective tissues as diverse as the paraspinal stromal tissues, ACL attachment regions to bone and synovial tissues of the ovine knee joint. Fibrillin-1 and elastin are both prominent components of elastic microfibrils and are considered to provide important compliance properties and elasticity to the intimal, medial and adventitial layers of developing and mature blood vessels (Carta et al. 2009; Kielty et al. 2002b; Rosenbloom et al. 1993; Rossi et al. 2010). These immunolocalization patterns were highly suggestive of important interactive properties between the perlecan and elastin. Interactions between the fibrillin-1 monomers and perlecan have already been reported and shown to be important in basement membrane assembly processes (Tiedemann et al. 2005), cellular attachment (Bax et al. 2007; Hubmacher et al. 2006; Jordan et al. 2006; Jovanovic et al. 2008; Midwood and Schwarzbauer 2002; Stephan et al. 2006; Williamson et al. 2007), developmental processes in blood vessels (Carta et al. 2009; Ramirez and Dietz 2007, 2009; Ramirez et al. 2007; Rossi et al. 2010), in the provision of viscoelastic properties to connective tissues (Kielty et al. 2002c; Midwood and Schwarzbauer 2002; Ramirez and Dietz 2007; Ramirez et al. 2007; Ritty et al. 2002; Stephan et al. 2006) and in cell-matrix attachments which are important in mechanotransduction (Hubmacher et al. 2006; Jordan et al. 2006; Midwood and Schwarzbauer 2002; Ramirez and Dietz 2007; Ritty et al. 2002; Stephan et al. 2006).

To better understand the potential interactions between the elastin and perlecan implied by their co-distributions in a number of connective tissues we employed the QCM-D technique which allows measurements of the thickness and compliance of the de novo assembled ECM deposited on to a quartz crystal surface. The QCM-D data presented in this study overcome an impediment to perlecan studies in that perlecan in different tissues can be subjected to different types of glycosylation and this may modulate the interactive properties of particular perlecan preparations. Endothelial cells synthesize an HS substituted form of perlecan (Whitelock 2002), while smooth muscle and synovial cells make a form decorated with both CS and HS side chains (Dodge et al. 1995). Perlecan is an intrinsic component of subendothelial and smooth muscle cell basement membranes where it binds fibrillin-1 and elastin influencing elastic fiber deposition (Kielty et al. 2007). Based on the distribution of perlecan and elastin observed in the present study, all of the aforementioned forms of perlecan interact with elastic fibers in situ. Association with perlecan presumably aids in the integration of elastic fibers within the surrounding extracellular matrix (Iozzo 1994, 2005; Melrose et al. 2008a; Whitelock et al. 2008). Endothelial cells exhibit strong integrin mediated attachment to fibrillin-1 and fibulin-5 coated surfaces; however they attach poorly to mature elastin, do not spread and have markedly impaired functional properties (Williamson et al. 2007). Thus a localization of perlecan with elastin would be expected to improve cell attachment in situ. Perlecan has cell attachment sites and acts as a linking module with a multitude of extracellular matrix components aiding in the integration of the elastic fibers into the surrounding extracellular matrix (Iozzo 1994, 2005; Melrose et al. 2008a; Whitelock et al. 2008). Vascular cells bind readily to perlecan; however, variations in the sub-domain structure of the HS chains and/or replacement of some of the HS chains with another glycosaminoglycan modulate its interactive capability (Whitelock et al. 1999; Whitelock and Iozzo 2005).

Fibrillin-1, is the major constituent of elastin microfibrils (Kielty et al. 2002b; Kielty et al. 2002c). It contains 47 epidermal growth factor domains and seven 8-cysteine-containing TB motifs (Kielty et al. 2002b; Kielty et al. 2002c; Midwood and Schwarzbauer 2002; Ramirez et al. 2007). Fibrillin-1 monomers are assembled head to toe into fibrillin-1 fibrils which support the coacervation of tropoelastin in situ and the microfibrillogenesis process (Bax et al. 2007), and as we have shown in the present study, the HS chains of perlecan located pericellularly also have interactive properties with tropoelastin which would be conducive to these processes. HS chains are also known to regulate N and C terminal interactions in fibrillin-1 (Cain et al. 2008) and in the retention of fibrillin-1 at the cell surface (Bax et al. 2007). Fibrillin-1 contains an Arg-Gly-Asp (RGD) sequence within the TB4 domain which interacts with cell surface integrins (Bax et al. 2007; Jovanovic et al. 2007; McGowan et al. 2008). These interactions have important roles to play in the microfibrillar assembly process and in cellular adhesion, αVβ3, α5β1 and αVβ6 integrins all interact with fibrillin-1 and display high (Kd 40 nM), moderate (Kd 450 nM) and low (Kd > 1 μM) binding affinities, respectively (Jovanovic et al. 2008), αvβ3 integrin regulates microfibril assembly in periodontal ligament cells (Tsuruga et al. 2009), and along with α5β1 integrin has a role in the activation of human umbilical vein endothelial cells (Mariko et al. 2010) and lung fibroblasts (McGowan et al. 2008), mediating cell signaling, proliferation, cell migration and in the provision of focal adhesions. These comments are consistent with fibrillin-1 fibrils which were observed in the present study, attached to cells in the posterior longitudinal ligament and outer annulus fibrosus and with regulatory roles in mechanotransductive processes which orchestrate extracellular matrix development and remodelling and maintenance of a homeostasic balance in extracellular matrix components. A genetic defiiency in fibrillin-1 results in pathological ageing of arteries in mice (Mariko et al. 2011). Elastic microfibrils have a particularly important role to play in arterial development and also provide compliance properties in mature vessels (Carta et al. 2009). Microfibrils come in direct contact with endothelial cells providing anchorage of these cells to the internal elastic lamina with fibrillin-1 having an important role to play in cell adhesion and recruitment of cells during vessel development (Bax et al. 2007; Midwood and Schwarzbauer 2002; Williamson et al. 2007).

Observations in tensional and weight-bearing connective tissues have also established the roles for fibrillin microfibrils in the provision of viscoelastic properties (Ritty et al. 2002), in the anchorage of cells in the extracellular matrix (Bax et al. 2007; Jovanovic et al. 2008; Midwood and Schwarzbauer 2002) and in mechanosensory processes important to tissue function (Kielty et al. 2002a; Kielty et al. 2005; Kielty et al. 2002c; Sherratt et al. 2003), and tissue homeostasis and remodelling in health and disease (Hubmacher et al. 2006; Jordan et al. 2006; Jovanovic et al. 2008; Kielty et al. 2002a; Kielty et al. 2002c; Ramirez et al. 2008; Ramirez and Dietz 2007; Ramirez et al. 2007).

In QCM-D experiments performed in this study, the complex formed between the tropoelastin and perlecan could be disassembled using 1 M NaCl, indicating that these complexes were formed through electrostatic interactions. The Voigt model provided an estimate on the thickness of the adsorbed material on the crystal surface (Voinova et al. 1999). Use of a truncated form of tropoelastin devoid of its C-terminal region demonstrated the important role of the C-terminus in binding to HS-perlecan, and identified perlecan as the likely donor of HS chains (Akhtar et al. 2011). The protein core of perlecan also mediated interactions with tropoelastin, so it is likely that this region also contributed to the assembly process between these protein moieties. These interactions did not involve electrostatic interactions but instead appeared to rely on hydrophobic interactions between the strategic regions of the tropoelastin molecule and the perlecan core protein. Similar interactions have already been described for lipoproteins with the perlecan core protein (Fuki et al. 2000; Pillarisetti 2000; Yamazaki et al. 2004). Indeed, perlecan domain II displays a strong homology with the low density lipoprotein receptor and may be involved in lipid clearance from the circulatory system (Whitelock et al. 2008).

In this study we found that the elastic fibers containing elastin and fibrillin-1 were prominent components of the developing outer annulus fibrosus and of the posterior longitudinal ligaments of the human fetal spine at 14 weeks gestational age. At this stage of development, the human fetal spine is predominantly a cartilaginous structure, as the vertebral bodies have yet to undergo ossification and the stratification of the outer annular lamellae of the intervertebral disc are only just becoming discernable. The elastin observed in the outer annulus fibrosus attaches the collagenous lamellar layers to the adjacent cartilaginous vertebral body rudiments (Hayes et al. 2011). These elastin and fibrillin-1 containing elastic fibers may represent tensioning devices whereby forces are applied to the developing annulus generating hoop-stresses which regulate the subsequent development of the annular arcades, which are the characteristic features of this tissue. We also observed elastin in the annulus fibrosus of the newborn ovine intervertebral disc as punctate amorphous deposits similar to those depicted in elastogenesis studies (Broekelmann et al. 2005; Czirok et al. 2006; Kozel et al. 2006). These deposits always occurred in close proximity to perlecan around the annular fibrochondrocytes and in some cases appeared to be associated with funnel-like projections that emanated from the annular fibrochondrocyte pericellular matrix. In the context of the newborn annulus fibrosus nascent elastogenesis, interactions presumably occur at the cell surface where crosslinking of tropoelastin forms insoluble elastin. Irreversible conversion to insolubility is facilitated by lysyl oxidase, which also forms part of the elastic microfibrillogenesis assembly machinery in situ (Czirok et al. 2006; Kozel et al. 2006; Wagenseil and Mecham 2007). In the present study we also observed fibrillin-1 fibrils emanating from pericellular perlecan in the annulus fibrosus; fibrillin-1 is essential for elastic microfibrillogenesis.

Perlecan and elastin were also co-localized in the ACL attachment regions in ovine stifle joints. Tendons consist mainly of type I collagen in a hydrated proteoglycan matrix with collagen accounting for 65–80% of the dry mass of this tissue, while elastin is a minor component (Cetta et al. 1982; Kannus 2000). Elastic fibers are found in the Achilles tendons of young and old rabbits (Ippolito et al. 1980), rats (Greenlee and Pike 1971), the fibrocartilage, mineralized fibrocartilage zones of the osteotendinous junction (Kannus 2000), insertion points of rat flexor tendons (Ritty et al. 2002) and now in ACL insertion sites as shown in the present study. The nuchal ligaments of adults have the highest elastin content of any human tissues (Cleary et al. 1967; Fukuda et al. 1984) which appears early in utero in the posterior longitudinal ligaments as demonstrated in the present study.

We can find only one previous study which has localized elastin with perlecan in basal laminar synovial structures or in blood vessels. This study visualized the elastic basal lamina in synovium using TEM and showed that HS was present pericellularly around synovial cells but actually not on the elastic lamina (Nagaoka et al. 2001). Based on some of the regenerative strategies under evaluation in vascular biology (Kielty et al. 2007; Stephan et al. 2006), perlecan and elastin are the two components that may have value in combination in tissue regenerative procedures including replacement blood vessels (Kielty et al. 2007; Koria et al. 2011; Lim et al. 2008). The present study is the first demonstration of the colocalization of elastin, and perlecan in connective tissues, and the binding of tropoelastin to the extracellular matrix/basement membrane, proteoglycans such as perlecan.

Acknowledgments

This study was funded by NHMRC Project Grant 512167 to JM and JW, Grant support was also received by ASW from the Australian Research Council and the National Health and Medical Research Council.

Copyright information

© Springer-Verlag 2011