Email: email@example.com. Present address: Faculty of Dentistry, University of Toronto, 124 Edward St. Toronto, Ontario, Canada M5G 1G6.
Reason for performing study: The role of matrix metalloproteases (MMPs) and the timeline of proteolysis during laminitis development are incompletely understood.
Objectives: To determine the temporal progression of selected MMPs and protease regulators during laminitis development.
Methods: Five clinically normal Standardbred horses received, via nasogastric intubation, an oligofructose (OF) bolus (10 g/kg bwt). Laminitis induction proceeded for 48 h followed by euthanasia. Lamellar biopsies were obtained prior to dosing and at intervals during the treatment period for analysis (12, 18, 24, 30 and 36 h and at 48 h following euthanasia). Tissue samples were analysed by real-time PCR, zymography and western blotting.
Results: Activation of proMMP-2 occurs either simultaneously or at least 12 h following lamellar basement membrane (BM) damage, while no activation of proMMP-9 is seen during OF laminitis induction. Aggrecanase gene expression increased initially at 12–18 h post OF dosing, similar to BM changes. Gene expression of TIMP-2, a MMP regulator, decreases during laminitis development.
Conclusions: The MMP-2/MT1-MMP complex may not play a major role in initiating lamellar BM damage. Aggrecanase and TIMP-2 gene expression appear related to BM lamellar changes.
Potential relevance: MMPs, historically thought to cause laminitis, do not appear to play an initiating role in the lamellar lesion. Other host derived proteases and degradation of alternative lamellar matrix components need to be considered.
Laminitis compromises the lamellar suspensory apparatus of the distal phalanx (Pollitt 2010). There is evidence that the lamellar basement membrane (BM) is the histological site of the primary laminitis lesion (Pollitt 1996; Pollitt and Daradka 1998), but the molecular mechanism and progression of events is not completely elucidated.
A major question of laminitis pathogenesis is the role of proteases in lamellar BM degradation and dysadhesion. The matrix metalloprotease (MMP) family is comprised of at least 24 members, which degrade a variety of extracellular matrix (ECM) components including BM laminins (Ln) and collagens as well non-ECM molecules (Chakraborti et al. 2003). The gelatinases, MMP-2 and MMP-9 as well as the membrane bound MT1-MMP are increased during the acute phase of carbohydrate laminitis (Johnson et al. 1998; Pollitt et al. 1998; Kyaw-Tanner and Pollitt 2004; Kyaw-Tanner et al. 2008). However, the factor(s) responsible for activation of these and possibly other MMPs during laminitis development is yet to be determined. A second metalloprotease family is the A disintegrin-like and metalloprotease with thrombospondin motifs (ADAMTS) family. These secreted proteases consist of at least 20 members, of which ADAMTS-4 (aggrecanase) degrades the proteoglycan of cartilage matrix (Tortorella et al. 1999; Jones and Riley 2005). Equine ADAMTS-4 has recently been cloned and found to be upregulated in both natural cases and laminitis induction models (Budak et al. 2009; Coyne et al. 2009). Proteases produced by inflammatory cells also play roles in ECM degradation. MMP-9 is produced in tertiary granules of neutrophils (Cowland and Borregaard 1999), while the serine proteases neutrophil elastase and cathepsin are present in primary granules (Owen 2008). While MMP-9 has long been associated with laminitis (Johnson et al. 1998; Pollitt et al. 1998), neutrophil elastase has also been shown recently to increase (de la Rebiere de Pouyade et al. 2009).
Control of protease expression and activation is normally tightly regulated, being expressed only as required in processes such as tissue development and remodelling. However, many pathological conditions in both human and veterinary medicine such as cancer, arthritis and fibrosis, result from an imbalance between MMP activation and inhibition (Clutterbuck et al. 2008). MMPs are regulated by transcriptional activation or repression, proenzyme activation by removal of the propeptide and inhibition by tissue inhibitors of MMPs (TIMPs) (Sternlicht and Werb 2001; Visse and Nagase 2003; Clark et al. 2008; Kessenbrock et al. 2010).
Our group has established that loss of the BM components, Ln-332 and collagen type IV, as well as changes specific to the Ln-332 γ2 subunit, occur early in the lamellar BM during carbohydrate laminitis induction and prior to clinical signs of lameness (Visser and Pollitt 2011). The aim of this study was to correlate the progression of protease activation and proteaseregulator expression with previously observed BM degradation events in the same tissue samples during carbohydrate-induced laminitis (Visser and Pollitt 2011).
Materials and methods
Experimental model and tissue collection
Laminitis was experimentally induced in normal Standardbred horses (age range 5–9 years) by alimentary carbohydrate overload using oligofructose (OF) (10 g/kg bwt) (van Eps and Pollitt 2006). Hoof lamellar biopsies (approximately 10 × 10 mm square) were collected prior to OF dosing and at 12, 18, 24, 30 and 36 h post OF dosing (POD), with the horse standing, sedated with the appropriate digital nerve blocked with local anaesthetic (Croser and Pollitt 2006). At 48 h POD, following euthanasia by overdose of barbiturate, lamellar samples were also obtained below the lateral and central biopsy sites. Experiments were conducted according to the animal ethics guidelines set by The University of Queensland Animal Ethics Committee. All animals were continuously monitored and inspected by the Consultant Veterinary Officer to the Animal Welfare Unit at the University of Queensland. If pain could not be alleviated promptly, early euthanasia, even if this was prior to the planned conclusion of the experiment, was a proviso of the protocol. For protein extraction and RNA isolation, tissue samples were flash frozen in liquid nitrogen and stored at -80°C.
Real time PCR
Total RNA was isolated from frozen hoof tissue using Trizol1. Tissue samples were homogenised in Trizol reagent1 (1 ml/100 mg tissue) using a power homogeniser2. RNA samples were treated with RNAse free DNAse3 to remove genomic DNA prior to cDNA synthesis. cDNA was synthesised from 1 µg of total RNA using the reverse transcription system3 with random hexamer primers3.
Gene expression in tissue samples was measured using real-time PCR analysis. Primer sequences were designed to equine-specific sequences using Primer Express software4. Oligonucleotides were synthesised by Sigma Genosys5. Primer sequences, concentrations and amplicon sizes are listed in Table 1. PCR reactions were performed in 10 µl reaction volumes in 384 well plates prepared using an epMotion 5075 robot6. PCR reactions consisted of 5 µl SYBR Green Master Mix4, 2 µl cDNA (1:5), 2 µl primer mix and 1 µl H2O. Reactions were carried out in triplicate using as AB7900 thermocycler4. Reaction conditions consisted of a initial denaturation step of 95°C for 10 min, followed by 50 cycles of denaturation 95°C for 15 s and elongation of 60°C for 1 min. Following amplification, melting curve analysis was subsequently performed by one cycle of 95°C, 2 min, 60°C, 15 s and 95°C, 15 s. Relative gene expression levels, normalised to β-2 microglobulin, were determined using the comparative Ct method (2−ΔΔCt) (Livak and Schmittgen 2001). PCR products were analysed by standard agarose gel electrophoresis (Sambrook et al. 1989) to confirm the expected size of PCR product. Samples were electrophoresed on 2% agarose gels embedded with 0.5 µg/ml ethidium bromide followed by analysis by UV illumination.
Table 1. Primer sequences used for real time PCR gene expression
Amplicon size (bp)
Primer concentration (nmol/l)
SDS-PAGE and immol/lunoblotting
To prepare total protein extracts, lamellar tissue was ground to a powder under dry ice using a mortar and pestle, followed by homogenisation on ice using a power homogeniser2 in RIPA buffer (20 mmol/l Tris-HCl pH 7.5, 150 mmol/l NaCl, 1% Triton-X 100, 0.5% sodium deoxycholate, 0.1% SDS containing Complete Protease Inhibitors7 1 ml/100 mg tissue, 5 × 20 s). The cell suspensions were incubated on ice for 30 min with occasional vortexing. Samples were centrifuged 10,000 g, 10 min at 4°C and protein concentration of the supernatants determined using the BCA protein assay8. To determine levels of MT1-MMP expression and processing in lamellar samples, 100 µg of total protein was separated by SDS-PAGE. Gels were electro-transferred onto polyvinyl-D-fluoride (PVDF) Hybond-P membrane9 at 30 V for 16 h at 4°C. Membranes were blocked and washed in milk solution (20 mmol/l Tris-HCl pH 7.4, 150 mmol/l NaCl, 0.1% Tween-20, 5% skim milk powder) prior to incubation in MT1-MMP (clone 113 5B7) primary antibody10 overnight at 4°C. Membrane washing was repeated followed by incubation with an HRP-conjugated secondary antibody11 for 1 h at room temperature. After washing in PBS, antibody was detected by chemiluminescence using the Super Signal substrate8 as per the manufacturer's instructions. Optimal exposures were obtained by exposure to x-ray film and bands were scanned using Adobe Photoshop and analysed by densitometry using ImageJ 1.38.
To determine MMP substrate activity and visualise protease forms present in lamellar samples, substrate zymography was performed. Gelatin zymography was performed as in Mungall and Pollitt (1999) with some modification. Total protein samples (20 µg) were run on 10% SDS-PAGE gels containing 0.1% gelatin12 under nonreducing conditions, 90 min at 110 V. Commercial human MMP-210 and equine buffy coat samples (MMP-9) were used as standards. Gels were incubated twice in 2.5% Triton for 30 min to remove SDS, followed by incubation overnight at 37°C in zymography incubation buffer (50 mmol/l Tris-HCl pH 7.5, 5 mmol/l CaCl2). Gels were stained 1 h with Coomassie Brilliant Blue (30% methanol, 10% acetic acid containing 0.5% w/v Coomassie Brilliant Blue) followed by destaining (30% methanol, 10% acetic acid) until clear bands are visible on blue background. Casein zymography was performed as for gelatin zymography except gels were embedded with 0.1% casein12. Casein gels were prerun at 110 V for 20 min prior to loading of samples. Gels were dried and scanned using Adobe Photoshop and analysed by densitometry using ImageJ 1.38.
Group means (n = 5) of each time point compared to 0 h samples were analysed by ANOVA with the Dunnett post t test13. Significance was set at P<0.05. Graphs represent the mean ± s.e.
Clinical and histological findings
All horses exhibited decreased appetite by 8 h POD and developed diarrhoea by 16 h. Digital pulses were initially detected at 13–29 h POD and all horses developed clinical signs of lameness, characterised by weight shifting, between 20–43 h POD (Croser and Pollitt 2006; Milinovich et al. 2007). Lameness was graded using the Obel grading scale (Obel 1948). Prior to euthanasia, Obel grades were as follows: Grade 1= one horse, Grade 2= one horse, Grade 3= 3 horses. Histological changes were observed by haematoxylin and eosin stain between 12–30 h POD (Croser and Pollitt 2006), while immunohistochemical changes in BM components were first observed at 12 h POD (Visser and Pollitt 2011).
Gelatinase expression during laminitis development
A representative gelatin zymogram from Horse 4 is shown in Figure 1a. Gelatin zymograms of lamellar tissue from 0 h biopsies contain small amounts of 92 kDa proMMP-9, 72 kDa proMMP-2 and 62 kDa active MMP-2 (Fig 1a). Increased expression of proMMP-9 was observed initially at 18–24 h POD in individual horses, however no processing to the active 82-kDa MMP-9 form was observed at any time (Figs 1a,b). Overall, the relative level of MMP-9 gene expression reflected protein levels, as gene expression increased as early as 12 h POD; however, significance was only observed at 24 h POD and later (Fig 1e).
Increased proMMP-2 expression was observed in lamellar biopsy tissues at 12–24 h in all horses (Figs 1a,c). Conversion of proMMP-2 to the active MMP-2 form occurred at 18–48 h POD (Figs 1a,d). Relative levels of MMP-2 gene expression did not reflect protein levels, as there was no significant increase (Fig 1f).
MT1-MMP expression during laminitis development
Lamellar tissue obtained from samples prior to OF dosing showed expression of proMT1-MMP. Increased levels of proMT1-MMP were observed only at 48 h POD; however, no processing to the active MT1-MMP form was observed (Fig 2a). Similarly, relative MT1-MMP gene expression levels did not show any significant change (Fig 2b).
ADAMTS-4 expression during laminitis induction
Relative levels of ADAMTS-4 gene expression were increased at 12 h POD in 3 of 5 horses compared to time 0. All horses showed significantly increased levels of gene expression at 18–48 h POD dosing (Fig 3).
TIMP expression during laminitis induction
Expression levels of the TIMP-2 gene were observed to decrease as early as 12 h POD in some horses, although significance was only observed from 30 h onwards (Fig 4a). Expression of TIMP-3 was variable among the horses throughout the biopsy timeline and thus no consistent trend in the gene expression can be confirmed (Fig 4b).
Laminitis pathogenesis may involve dysregulation of MMP activity, a premise that links protease activity to lamellar basement membrane destruction and dysadhesion. It was inferred from increased MMP expression associated with laminitis, that these proteases were the causative agents of lamellar damage (Johnson et al. 1998; Pollitt et al. 1998; Kyaw-Tanner and Pollitt 2004; Kyaw-Tanner et al. 2008; Loftus et al. 2008, 2009). However, these studies involved single time point analyses, usually after clinical signs were manifest. Using our novel technique of lamellar biopsy during OF induction (Croser and Pollitt 2006; Visser and Pollitt 2011) we have determined that BM changes, including loss of collagen type IV and Ln-332, occur as early as 12 h POD, well before the appearance of laminitis clinical signs (Visser and Pollitt 2011). Here we report (using the same tissue samples) that conversion of proMMP-2 to its active form occurs simultaneously with or up to 12 h after detection of BM changes. Thus, we conclude that MMP-2 may not be the protease responsible for initiating the BM damage observed in carbohydrate overload laminitis.
Although MMP-2 may not be an initiating factor in laminitis, a role in intensifying BM damage cannot be ruled out. Thus, the mechanism of how this MMP is regulated during laminitis is still important to consider. One of the primary mechanisms of MMP regulation is at the level of transcription; however, in this study no increase in the level of MMP-2 gene transcription is observed. This is contrary to previous observations by our group (Kyaw-Tanner and Pollitt 2004) but similar to results obtained from samples at various time points induced with black walnut extract (BWE) (Loftus et al. 2006, 2007).
Similarly, there was no change in MT1-MMP transcription or activation. This further contradicts the notion of lamellar MMP-2 involvement and activation through the MT1-MMP/ TIMP-2 complex. A recent study also found that MT1-MMP gene expression itself as well as the MT1-MMP activator PACE4 were not significantly upregulated in either carbohydrate overload induction or chronic laminitis samples (Loftus et al. 2009). Additionally, decreased TIMP-2 gene expression was observed as early as 12 h POD. Thus our TIMP-2 gene expression data coincide with those of others (Kyaw-Tanner et al. 2008; Coyne et al. 2009) and add that TIMP-2 repression decreases as laminitis progresses. TIMP-2 is able to bind to and inhibit all MT-MMPs (Lambert et al. 2004). If TIMP-2 protein levels follow gene expression during laminitis development, as observed here, this may allow for uncontrolled activation of other BM degrading MT-MMPs and subsequent BM damage. Dysregulation of the MMP/TIMP balance during laminitis appears specific for TIMP-2 as there was no consistent change in gene expression of TIMP-3, similar to other reports (Coyne et al. 2009). Thus, either direct cleavage of BM components by MMP-2 or MT1-MMP or by subsequent protease activation relays are unlikely to be involved in early disease progression.
The ADAMTS proteases are involved in cartilage matrix degradation (Jones and Riley 2005) by specifically cleaving the chondroitin sulphate proteoglycan, aggrecan (Tortorella et al. 1999; Nakamura et al. 2000). A role for ADAMTS-4 in laminitis development has recently been suggested (Budak et al. 2009; Coyne et al. 2009). In this study ADAMTS-4 expression was increased at the first sampling point in some horses, suggesting an early role for this molecule. Attempts to evaluate ADAMTS-4 protein levels in lamellar extracts using commercial antibodies were unsuccessful in our laboratory, but active protein expression as well as release of aggrecan or other proteoglycan fragments in laminitis development still needs to be confirmed.
Limited proteolysis of collagen type IV by ADAMTS-4 does occur (Lauer-Fields et al. 2007). We have shown that lamellar BM collagen type IV is lost as early as 12 h POD (Visser and Pollitt 2011), the same time that ADAMTS-4 gene expression begins to increase in some horses. In addition to aggrecan, ADAMTS-4 cleaves other chondroitin sulphate proteoglycans such as brevican and versican (Nakamura et al. 2000; Westling et al. 2004) as well as members of the small leucine rich proteoglycan group, decorin and biglycan (Kashiwagi et al. 2001). Similar to cartilage, large polysulphated proteoglycans are present in lamellar ECM (Pawlak et al. 2009; Wang et al. 2009b) and their loss from the lamellar BM during laminitis development may compromise lamellar suspensory function (Pollitt 2010).
Increased lamellar zymogen MMP-9 as well as increased MMP-9 gene transcription concurs with other reports (Johnson et al. 1998; Black et al. 2006; Loftus et al. 2006, 2008, 2009; Noschka et al. 2009). The accumulation of proMMP-9 in the lamellae occurs early during laminitis induction, at 18–24 h. However, no conversion to the active form was observed so it is unlikely that this MMP has a causative role in lamellar damage or downstream protease activation.
In this study we have shown that the proteases MMP-2 and MMP-9, historically associated with laminitis development, appear not to initiate laminitis pathogenesis. Rather, other proteases such as ADAMTS-4 and regulators such as TIMP-2 may be involved in early lamellar damage. Alternative proteases and the degradation of additional lamellar matrix components need to be considered in further studies.
Authors' declaration of interests
No conflicts of interest have been declared.
Source of funding
Funding was provided by the Rural Industries Research and Development Corporation of Australia and the Animal Health Foundation of St Louis, Missouri, USA. M.B.V. was the recipient of an Endeavour International Postgraduate Research Scholarship from the University of Queensland.
We thank Emma Croser, Christopher Owens, Katie Asplin, Ali Nourian and Marianne Keller for assistance with tissue collection and animal experiments.
1 Invitrogen, Carlsbad, California, USA.
2 Omni International Inc, Marietta, Georgia, USA.
3 Promega, Madison, Wisconsin, USA.
4 Applied Biosystems, Foster City, California, USA.
5 Sigma Genosys, Castle Hill, New South Wales, Australia.
6 Eppendorf, Hamburg, Germany.
7 Roche, Basel, Switzerland.
8 Pierce, Rockford, Illinois, USA.
9 GE Healthcare, Piscataway, New Jersey, USA.
10 Calbiochem, Gibbstown, New Jersey, USA.
11 Zymed, South San Francisco, California, USA.
12 Sigma, St. Louis, Missouri, USA.
13 GraphPad Software, La Jolla, California, USA.
Author contributions M.B.V. designed and performed experiments and prepared the manuscript. C.C.P. designed experiments and edited the manuscript.