Drs. Lim and Meinjohanns contributed equally to this work.
In Vivo Imaging of Matrix Metalloproteinase 12 and Matrix Metalloproteinase 13 Activities in the Mouse Model of Collagen-Induced Arthritis
Version of Record online: 25 FEB 2014
Copyright © 2014 by the American College of Rheumatology
Arthritis & Rheumatology
Volume 66, Issue 3, pages 589–598, March 2014
How to Cite
Lim, N. H., Meinjohanns, E., Bou-Gharios, G., Gompels, L. L., Nuti, E., Rossello, A., Devel, L., Dive, V., Meldal, M. and Nagase, H. (2014), In Vivo Imaging of Matrix Metalloproteinase 12 and Matrix Metalloproteinase 13 Activities in the Mouse Model of Collagen-Induced Arthritis. Arthritis & Rheumatology, 66: 589–598. doi: 10.1002/art.38295
- Issue online: 25 FEB 2014
- Version of Record online: 25 FEB 2014
- Accepted manuscript online: 27 NOV 2013 09:11AM EST
- Manuscript Accepted: 21 NOV 2013
- Manuscript Received: 24 JUN 2013
- Arthritis Research UK
- European Union Sixth Framework Programme (project CAMP)
- Seventh Framework Programme (project LIVIMODE)
- NIH (National Institute of Arthritis and Musculoskeletal and Skin Diseases. Grant Number: AR-40994
- Ministero dell'Istruzione dell'Università e della Ricerca. Grant Number: MIUR grant Prin 2007
To develop enzyme-activatable Förster resonance energy transfer (FRET) substrate probes to detect matrix metalloproteinase 12 (MMP-12) and MMP-13 activities in vivo in mouse models of inflammatory arthritis.
Peptidic FRET probes activated by MMP-12 and MMP-13 were reverse designed from inhibitors selected from a phosphinic peptide inhibitor library. Selectivity of the probes was demonstrated in vitro using MMP-1, MMP-2, MMP-3, MMP-12, and MMP-13. In vivo activation of the probes was tested in the zymosan-induced mouse model of inflammation, and probe specificity was evaluated by the MMP inhibitor GM6001 and specific synthetic inhibitors of MMP-12 and MMP-13. The probes were used to monitor these enzyme activities in the collagen-induced arthritis (CIA) model in vivo.
The MMP-12 and MMP-13 activity probes (MMP12ap and MMP13ap, respectively) discriminated between the activities of the 2 enzymes. The in vivo activation of these probes was inhibited by GM6001 and by their respective specific inhibitors. In the CIA model, MMP12ap activation peaked 5 days after disease onset and showed strong correlation with disease severity during this time (r = 0.85, P < 0.0001). MMP13ap activation increased gradually after disease onset and correlated with disease severity over a longer period of 15 days (r = 0.58, P < 0.0001).
We generated two selective FRET probes that can be used to monitor MMP-12 and MMP-13 activities in live animals. MMP12ap follows the initial stage of inflammation in CIA, while MMP13ap follows the progression of the disease. The specificity of these probes is useful in monitoring the efficacy of MMP inhibitors.
Rheumatoid arthritis (RA) is a chronic systemic autoimmune disease that mainly affects joints. It is characterized by synovial inflammation, with infiltration of immune cells and hyperplasia of synovial cells, followed by articular cartilage destruction and joint dysfunction. The matrix metalloproteinases (MMPs), a family of zinc-dependent metalloendopeptidases, many of which are capable of degrading components of the extracellular matrix (), are involved in multiple stages of the progression of RA (). Detection of the different MMP activities in vivo may offer novel ways of monitoring the pathogenesis of RA in the clinic.
Early in the pathogenesis of RA, there is an increase in the number of macrophages that infiltrate the synovium, and the extent of infiltration correlates with disease severity (). A unique MMP associated with macrophages is MMP-12, or macrophage metalloelastase (). Concurrent with the increase in macrophages in the synovium, there is an increased expression of active MMP-12 associated with these macrophages in RA patients (). Transgenic rabbits overexpressing MMP-12 showed increased synovial hyperplasia, pannus formation, and articular cartilage degeneration in inflammatory arthritis (). Detection of MMP-12 activity could be a surrogate marker for macrophage infiltration in RA. Articular cartilage destruction, in contrast, occurs later in the pathogenesis of RA, and several MMPs are involved in the degradation of type II collagen in articular cartilage (). In mice, the major collagen-degrading enzyme is MMP-13, or collagenase 3. Inhibition of MMP-13 decreased erosion in models of RA (), and mice lacking MMP-13 were protected against cartilage damage in an osteoarthritis model (). Detection of MMP-13 activity would be a marker for tissue destruction in RA, at least in murine models of arthritis.
The activities of these MMPs in disease pathology have so far been inferred from limited snapshots which are affected by several aspects that govern activity, including transcription, posttranslational modification, endocytosis, activation of the MMPs, and their inhibition by endogenous inhibitors ([9-11]). In vivo optical imaging allows the effect of the sum of these processes on the activities of the MMPs to be measured in real time through the use of near infrared (NIR) fluorophores, since fluorescence is detected in the NIR window of the spectrum, where tissue is largely transparent, and may be detected up to a depth of ∼1 cm (). This makes NIR fluorophore optical imaging suitable for small animal imaging and for clinical imaging of the joints of the hands and feet in humans. We chose to develop probes specific for MMP-12 and MMP-13 to detect the activities of these enzymes during the progression of RA, since they represent macrophage migration and cartilage collagen degradation.
There are several designs of probes for detecting protease activity in vivo (). Activity-based probes are chemical compounds that act as an irreversible inhibitor of the protease by covalently modifying their targets with the fluorophore. Some of these have been developed for cysteine proteases ([14-16]). However, detection of low abundance proteases by this approach is limited, as these probes do not multiply signals through catalysis. Substrate-based probes, on the other hand, overcome this limitation by amplifying the signal from a single enzyme. These consist of peptides that have an NIR fluorophore attached to one side of the proteolytic cleavage site and a Förster resonance energy transfer (FRET) quencher on the other. These have been developed for cysteine proteases ([17-19]), serine proteases (), and metalloproteinases (). A drawback of using substrate-based probes is that the products do not necessarily stay in the vicinity of the enzyme. To overcome this limitation, peptide substrates flanked on one side by a cationic cell–penetrating peptide with an NIR fluorophore and on the other by an anionic sequence with a quencher have been developed (). Upon proteolysis, the cleaved cationic fragment with the NIR fluorophore interacts with the lipid membrane and penetrates into the cell. Another method of limiting product diffusion is to incorporate multiple substrates on large polymeric scaffolds, which have been developed to detect a broad range of MMPs () and cathepsins (). Recent developments for substrate-based probes include the reverse design of specific tight-binding inhibitors into peptide substrates of cysteine cathepsins ().
To develop selective MMP-12 and MMP-13 probes, we screened a phosphinic peptide inhibitor library for high-affinity inhibitors and reverse designed the inhibitors into selective FRET peptide substrates for these MMPs. The specificity of these substrates for MMP-12 and MMP-13 were confirmed in vitro and in vivo. These substrates, the MMP activity probes MMP12ap and MMP13ap, were then used to characterize MMP-12 and MMP-13 activities in the collagen-induced arthritis (CIA) model of RA. The combination of the two probes demonstrates the longitudinal changes in the activity of these two distinct MMPs during the progression of the disease.
MATERIALS AND METHODS
Synthesis and screening of inhibitor libraries
The synthesis of the phosphinic peptide inhibitor library and its screening method have been described previously (). Briefly, the solid-phase combinatorial library of ∼500,000 phosphinic peptides was constructed on polyethylene glycol polyacrylamide (PEGA) resin using a one-bead-two-compounds approach, where each bead contained a general FRET MMP substrate and a putative phosphonic peptide inhibitor. Upon incubation with MMP-12 or MMP-13, beads containing a weak inhibitor fluoresce, whereas beads containing a strong inhibitor remain dark. The inhibitor sequence on the dark beads was determined by matrix-assisted laser desorption ionization–time-of-flight mass spectrometry, and the identified sequence was used to generate peptide substrates for MMP-12 and MMP-13.
MMP-12 and MMP-13 FRET peptide substrate synthesis
Peptides were synthesized by multiple-column peptide synthesis on PEGA800 resin in a 20-column synthesis block (). Briefly, 9-fluorenylmethoxycarbonyl (FMOC)–Gly-OH was coupled to PEGA800 resin (0.2 mM/gm) in an o-(benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium tetrafluoroborate (TBTU) (2.8 equivalents) and N-ethylmaleimide (NEM) (4 equivalents) coupling step. FMOC was removed by treatment with 20% piperidine in dimethylformamide (DMF). The FRET substrates were synthesized by coupling after preactivation of Nα-FMOC amino acids (3 equivalents) with TBTU (3 equivalents) and NEM (4 equivalents) using the following side chain–protecting groups for the Nα-FMOC amino acids: OtBu for Asp and Glu; tBu for Tyr, Ser, and Thr; Trt for Cys, Asn, and Gln; Boc for His, Lys, and Trp; and Pmc for Arg. Each coupling step was carried out for 3 hours using 3 equivalents of the preactivated amino acid in 0.5 volumes of DMF. In difficult couplings, 0.2 equivalents of 3,4-dihydro-3-hydroxy-4-oxo-1,2,3-benzotriazine was added as an acylation catalyst and indicator of reaction completeness. Reaction completion was further assessed using the Kaiser test. Two equivalents of the preactivated amino acid were added to coupling reactions that were incomplete. The terminal amino protecting FMOC was removed by piperidine treatment. Peptides were cleaved from the resin, and side chain deprotection groups were removed by treatment with trifluoroacetic acid (TFA):thioanisole:ethanedithiol:water (87.5:5:2.5:5). FRET substrates were purified by reverse-phase high-performance liquid chromatography (HPLC) with a DeltaPak C18 column and a linear gradient of 0–90% acetonitrile with water containing 0.1% TFA on a Waters HPLC system.
In vivo probe assembly
The in vivo probes were synthesized using the synthesis protocol described in the previous section, with the following modifications to incorporate the Cy5.5 fluorophore and QSY21 quencher. The base-labile linker hydroxymethyl benzoic acid was coupled to the PEGA800 resin (0.4 mM/gm) by TBTU (2.9 equivalents) and NEM (4 equivalents), with preactivation for 1 minute and coupling for 2 hours. The FMOC-Gly-OH amino acid (3 equivalents) was coupled by 1-(mesitylene-2-sulfonyl)-3-nitro-1,2,4-triazole (3 equivalents) and methylimidazole (6 equivalents), with preactivation for 1 minute and esterification for 2 hours. The first lysine was protected with the (4-methoxyphenyl)-diphenylmethyl (MMT) group. The terminal FMOC group was removed with 20% piperidine in DMF, and the protected peptide was cleaved from the resin by 0.1% NaOH. The QSY21 quencher was coupled for 12 hours to the terminal amino group by using QSY21-NHS ester (1.2 equivalents) in DMF and addition of N,N-diisopropylethylamine (DIPEA). The QSY21-labeled peptides were purified by semipreparative HPLC under neutral conditions before being subjected to moist 0.5% TFA in CH2Cl2 for 15 minutes to remove MMT. The Cy5.5 fluorophore was coupled to the side chain amino group of lysine by succinimide ester coupling using Cy5.5-OSu in DMF with DIPEA for 12 hours. The NIR fluorophore–labeled peptides were deprotected with 95% TFA in water for 2 hours.
Characterization of peptides
Peptides were purified by HPLC and analyzed by electrospray ionization mass spectrometry (ESI-MS) on a QTOF Waters Global Ultima system, using aqueous acetonitrile with 0.2% formic acid for the eluting spray. Individual products analyzed were as follows: MMP12-1, sequence and composition Y(NO2)GPLG∼LEEAK(Abz)G-NH2 (C58H86N15O19), retardation factor (Rf) 0.168 and theoretical mass/charge by MS(ESI) 1296.6224 [M+H]+, observed 1296.5868 [M+H]+; MMP12-2, sequence and composition Y(NO2)YIYG∼LTMPGK(Abz)G-NH2 (C72H101N16O19S), Rf 0.198 and theoretical mass/charge by MS(ESI) 1525.7150 [M+ H]+, observed 1525.7172 [M+H]+; MMP13-1, sequence and composition Y(NO2)GPLG∼MRGLK(Abz)G-NH2 (C58H89N17O16S), Rf 0.177 and theoretical mass/charge by MS(ESI) 1311.6394 [M+H]+, observed 1311.6251 [M+ H]+; and MMP13-2, sequence and composition Y(NO2)GPAG∼LYEK(Abz)G-NH2 (C56H75N13O18), Rf 0.154 and theoretical mass/charge by MS(ESI) 1216.5353 [M+H]+, observed 1216.5903 [M+H]+.
For MS analysis of the NIR fluorophore substrates, sodiated species of the sulfate groups in Cy5.5 were observed in various amounts. Due to two internal positive charges, species were observed as double and triple charged under acidic conditions. The most prevalent ions were as follows: MMP12ap, sequence and composition QSY21-GPLG∼LEEAK (Cy5.5)G-OH (C124H141N16Na3O32S5), Rf 0.104 and theoretical mass/charge by MS(ESI) 1298.8342 [M]++/2, observed 1298.8356 [M]++/2 and 873.8271[M+H]+++/3; and MMP13ap, sequence and composition QSY21-GGPAG∼LYEK(Cy5.5)G-OH (C124H138N16Na4O31S5), Rf 0.124 and theoretical mass/charge by MS(ESI) 1299.3949 [M]++/2, observed 1299.3801 [M]++/2, 1288.3898 (3 Na+), and 1277.4134 (2 Na+).
Catalytic domains of MMP-1 (MMP-1cat) and MMP-3 (MMP-3cat), full-length MMP-2, full-length MMP-12, and full-length MMP-13 were prepared as previously described ([26-28]). The MMP-12–specific inhibitor (MMP-12i), RXP470, is the phosphinic peptide inhibitor described as compound 1 by Devel et al (), and the MMP-13–specific inhibitor (MMP-13i), EN211, is the N-O-isopropyl sulfonamido–based hydroxamate described as compound 5 by Nuti et al ().
The rates of cleavage of the peptides by different MMPs were determined by incubation of 1 μM of the peptide (MMP12-1, MMP12-2, MMP13-1, or MMP13-2) with 1 nM of the enzyme (MMP-1cat, MMP-2, MMP-3cat, MMP-12, or MMP-13) in 50 mM Tris HCl (pH 7.5), 150 mM NaCl, 10 mM CaCl2, 0.02% NaN3, and 0.05% Brij-35 for 1 hour at 37°C. Steady-state cleavage of the substrates was continuously monitored by reading the emission at 393 nm following excitation at 325 nm using a fluorescence plate reader (SpectraMax; Molecular Devices). The fluorescence of completely cleaved substrate was used to convert fluorescence units into moles. The Km values of the substrates for MMP-12 and MMP-13 were determined by incubation of 1 nM of enzyme with a range of substrate concentrations (250 nM to 64 μM) for 30 minutes. The rates of substrate hydrolysis were calculated and the data were fit to the following equation to obtain the Km and maximal velocity (Vmax) using GraphPad Prism software version 5.0: v = (Vmax ×[S])/(Km + [S]). The Kcat value was derived from Kcat = Vmax/[E].
Male C57BL/6J mice and DBA/1 mice age 10 weeks were purchased from Harlan Laboratories. Mice were housed in groups of 6 in individually vented cages maintained at 21 ± 2°C on a 12-hour light/dark cycle with food and water provided ad libitum. All experimental protocols were performed in compliance with the UK Animals (Scientific Procedures) Act 1986 regulations for the handling and use of laboratory animals (PPL: 70/6533).
Paw edema was induced by injection of 20 μg of zymosan A in sterile phosphate buffered saline (PBS; Sigma-Aldrich) into the right hind footpad of 10-week-old male C57BL/6J mice following induction of anesthesia by gaseous anesthetic (2% isoflurane and O2). Two days after injection, probes were delivered either intraplantar (10 μl of 1 μM probe with or without inhibitors [10 μM GM6001, 1 μM MMP-12i, or 1 μM MMP-13i]) or intravenously via the tail vein (150 μl of 1 μM probe). The amount of probe injected was based on a compromise between detection, possible toxicity, and cost. A test injection of 5 times more probe intravenously yielded a corresponding 5-fold increase in signal with no observable side effects (data not shown). Two, 4, 8, and 16 hours after injection, fluorescence images of the mice were obtained by capturing the emission at 700 nm for 1 minute after exposure at 630 nm under gaseous anesthesia (2.5% isoflurane and O2) in a Kodak In Vivo FX Pro (Carestream).
For the CIA model, 10-week-old DBA/1 mice were immunized by intradermal injection of an emulsion of 200 μg of bovine type II collagen in 100 μl of Freund's complete adjuvant into the base of the tail. Arthritis was assessed daily by scoring each limb of the mice according to the method of Inglis et al (), where 0 = normal, 1 = slight swelling and erythema, 2 = pronounced edematous swelling, and 3 = joint rigidity. The thickness of each hind paw was also measured with microcalipers (Kroeplin). Three, 5, 10, or 15 days after the onset of arthritis, 150 μl of 1 μM probe was delivered intravenously via the tail vein. Four hours after injection, fluorescence images of the mice were obtained.
Image and statistical analysis
Fluorescence images were analyzed using the Carestream MI software (version 5.1). Regions of interest (ROIs) were defined using the wand tool to encompass the entire hind paw. The mean fluorescence intensity (MFI) of the ROI was obtained. Fluorescence images shown are false colored using the rainbow spectrum, with the range indicated in the figures. All statistical analysis was carried out using Prism software using paired Student's t-tests, one-way analysis of variance (ANOVA) with Bonferroni post tests, and Pearson's correlation where indicated.
Selection of FRET substrates by reverse design of inhibitors
Screening the phosphinic peptide inhibitor bead library with MMP-12 yielded the sequences GPLG[PO2H-CH2]LEEA and YIYG[PO2H-CH2] LTMPG as MMP-12 inhibitors. These were reverse designed and synthesized as the MMP12-1 (Y(NO2) GPLGLEEAK(Abz)G) and MMP12-2 (Y(NO2) YIYGLTMPGK(Abz)G) FRET substrates. Screening the same library with MMP-13 yielded the sequences GPLG[PO2H-CH2]MRGL and GGPAG[PO2H-CH2] LYEK. Similar reverse design led to the MMP13-1 (Y(NO2)GPLGMRGLK(Abz)G) and MMP13-2 (Y(NO2)GGPAGLYEK(Abz)G) FRET substrates.
In vitro activation of probes
We tested the activation of the 4 peptide substrates MMP12-1, MMP12-2, MMP13-1, and MMP13-2 in vitro by various MMPs. The results are summarized in Table 1. The MMP12-1 substrate was cleaved 10 times more rapidly by MMP-12 than by MMP-13 and 40–75 times more rapidly by MMP-12 than by MMP-1, MMP-2, or MMP-3. MMP12-2 was cleaved by MMP-12 30 times more rapidly than by all of the other MMPs tested. MMP13-1 was poorly cleaved by all of the MMPs tested, whereas MMP13-2 was cleaved 2 times more rapidly by MMP-13 than by MMP-12, and 10–20 times more rapidly by MMP-13 than by MMP-1, MMP-2, or MMP-3.
The kinetic parameters for the cleavage of the FRET substrates by MMP-12 and MMP-13 were determined. MMP12-1 was cleaved by MMP-12 with a Km of 91.51 μM and a Kcat of 4.51 seconds−1. MMP12-2 was cleaved by MMP-12 with a Km of 110.5 μM and a Kcat of 3.35 seconds−1. No concentration-dependent cleavage of MMP13-1 was detected, indicating that the inhibitor peptide was unsuitable as a substrate, whereas MMP13-2 was cleaved by MMP-13 with a Km of 58.4 μM and a Kcat of 4.34 seconds−1. The substrates MMP12-1 and MMP13-2 were selected for conversion into in vivo probes, and the fluorophore and quencher pair were switched to Cy5.5 and QSY21. These in vivo substrates will henceforth be referred to as MMP12ap for the MMP-12 activity probe and MMP13ap for the MMP-13 activity probe. Their chemical structures are shown in Figure 1A. Retention of the specificity of cleavage of these in vivo probes was qualitatively shown by incubation of 1 μM of the substrate with 50 nM of the respective MMPs for 30 minutes, and the reaction was spotted on a glass slide and imaged (Figure 1B).
In vivo activation of probes in acute inflammation
The zymosan-induced paw edema model of acute inflammation was used to test probe activation in vivo. After intraplantar injection of 150 μl of 1 μM probes 2 days after the zymosan injection, both MMP12ap and MMP13ap were found to be activated in a time-dependent manner. (Representative images from a single mouse are shown in Figure 2A.) Quantification of the MFI of the whole paw showed that the signal from MMP12ap peaked 8 hours after probe injection, at 11,600 units, whereas the signal from MMP13ap peaked 4 hours after injection, at 6,700 units (Figure 2B). Coinjection with the broad-spectrum MMP inhibitor GM6001 at 10 times the probe concentration inhibited the cleavage of these probes, indicating that MMPs were largely responsible for the observed increase in fluorescence.
Further validation of the specificity of probe activation in vivo was achieved by coinjection of the specific inhibitors MMP-12i and MMP-13i. Coinjection of 1 μM MMP-12i with 1 μM MMP12ap showed that it competed with the probe at the early time periods (0–2 hours) and achieved full inhibition of probe activation by 8 hours (Figure 2B), whereas coinjection of 1 μM MMP-13i did not have a significant effect on activation of MMP12ap. Activation of MMP13ap was effectively inhibited by coinjection with 1 μM MMP-13i or 10 μM GM6001 for 2 hours after injection (Figure 2B). However, this inhibition of probe activation was not maintained at later time points. This loss of inhibition could be due to enzyme turnover or clearance of inhibitor. MMP-12i had no effect on the activation of MMP13ap. These data indicate that MMP-12 and MMP-13 are mostly responsible for MMP12ap and MMP13ap activation, respectively, in the inflamed paws.
To apply the MMP12ap and MMP13ap probes to other systemic disease models, such as CIA, intravenous delivery of the probes was first tested in the zymosan model of rapid inflammation. Upon intravenous injection of 150 μl of 1 μM of either probe, there was a transient increase in fluorescence intensity over time in the inflamed paw for both probes, with maximum differentiation of the inflamed paw from the contralateral paw at 4 hours (Figure 2C). A maximum MFI of 6,000 units for MMP12ap and 13,300 units for MMP13ap was observed at 4 hours in the inflamed paw (Figure 2D). Increased fluorescence was detected in the contralateral paw, presumably due to circulation of activated probe (increased from a background of 500 units to 3,000–4,000 units for MMP12ap and 5,000–8,000 units for MMP13ap). However, the signal continued to be significantly higher in the inflamed paw 4 and 16 hours after injection of MMP12ap, and 4, 8, and 16 hours after injection of MMP13ap. This signal was 1.5 and 1.9 times higher at 4 hours in the inflamed paw than in the contralateral paw for MMP12ap and MMP13ap, respectively. These data indicate that probes delivered via the tail vein are activated mainly in inflamed paws.
In vivo activation of probes in the CIA model
In the CIA model, 68% of the animals (49 of 72) had at least 1 swollen paw 45 days after induction (Figure 3A). Figure 3B shows the total disease severity scores on days 3, 5, 10, and 15 after onset of inflammation in the animals that developed disease. As a measure of disease severity, hind paw thickness was consistent with the clinical score (Figure 3B). Typical images obtained using the MMP12ap and MMP13ap, together with their individual clinical scores, are shown in Figures 3C and D. Both the incidence of disease and the amount of signal obtained were lower for the fore paws than for the hind paws. We therefore excluded the data from the fore paws from this study.
The amount of MMP12ap activation in each inflamed hind paw 3, 5, 10, and 15 days after the onset of inflammation is shown in Figure 4A. The amount of probe activation correlated with the amount of swelling observed when the probe was delivered 3 and 5 days after the onset of disease (Pearson's correlation coefficient r = 0.991, P = 0.009 for day 3 and r = 0.759, P = 0.006 for day 5). There was no significant correlation 10 days after disease onset (r = −0.031, P = 0.953) or 15 days after disease onset (r = 0.488, P = 0.512). The correlation was strongest when the data from 3 days and 5 days after onset were taken together (r = 0.847, P < 0.0001), but became weaker when all time points were taken together (r = 0.416, P = 0.039). The correlation data are summarized in Table 2. This indicates that MMP12ap was most useful at classifying disease severity at an early window of up to 5 days after onset of disease. Activation was also highest 5 days after onset, with an MFI of 1,630 units (Figure 4C).
|Days after onset of arthritis|
|3||5||10||15||3 and 5||All|
Figure 4B shows MMP13ap activation in individual inflamed hind paws 3, 5, 10, and 15 days after the onset of inflammation. The correlation between MMP13ap activation and paw swelling is summarized in Table 2. Looking at the individual days, correlation between probe activation and paw swelling was significant only on day 5 (r = 0.833, P = 0.02). However, when all of the data were taken together, the correlation between MMP13ap activation and disease severity was significant (r = 0.561, P < 0.0001). This indicated that MMP13ap was better at classifying disease over a longer timeframe than MMP12ap. Activation of MMP13ap was highest 15 days after onset, with an MFI of 1,300 units (Figure 4D).
The versatility, specificity, and signal intensity of the in vivo probes presented here will allow the monitoring of joint damage throughout the development of arthritis. MMP12ap can be used to monitor the initial stages of inflammation, whereas MMP13ap can be used throughout disease progression.
We characterized two peptide sequences that are selectively cleaved by MMP-12 and MMP-13, respectively, through the reverse design of MMP-12– and MMP-13–specific phosphinic inhibitors. These sequences were synthesized as the NIR fluorophore FRET MMP12ap and MMP13ap substrate probes. Using these probes in live animals, we demonstrated that there is clear temporal regulation of MMP-12 and MMP-13 activity in the CIA model of arthritis, with an early peak in MMP-12 activity, followed by later MMP-13 activity. Additionally, activation of MMP12ap correlated strongly with early disease progression, and activation of MMP13ap correlated with disease severity over a longer period.
Immunohistochemistry has previously shown that there is increased expression of MMP-12 associated with the inflammatory macrophages lining the synovium in RA patients (). With MMP12ap, we visualized an increase in MMP-12 activity in the joints of live animals. This activity peaks 5 days after the onset of CIA and declines by 10 days. This is similar to that observed through immunohistochemistry in the rabbit model of RA (). MMP-12 may also activate other MMPs, such as MMP-2 and MMP-3 (). Indeed, MMP-3 activity, as detected by a polymeric MMP-3 substrate probe, peaked 5 weeks after immunization in an inflammatory model of arthritis (), suggesting that MMP-3 activation occurs after macrophage-derived MMP-12 activity. The eventual down-regulation of MMP-12 activity after macrophage infiltration may serve to limit proinflammatory cytokine processing by MMP-12 (). MMP12ap is therefore an early marker of the infiltration of activated macrophages into tissue. Detection of macrophage infiltration using positron emission tomography has been suggested as a means to diagnose arthritis before clinical symptoms manifest (). MMP12ap, on the other hand, serves as a cheaper, less damaging method to detect early macrophage infiltration.
MMP-13 is a terminal effector of cartilage degradation due to its ability to degrade the type II collagen component of cartilage (). The breakdown products of type II collagen can be monitored in the serum of RA patients and have been shown to correlate with the amount of radiologic damage (). Overexpression of MMP-13 in mice results in degradation of articular cartilage (). Gene expression profiling of the paws of mice with CIA also indicated that MMP-13 was up-regulated at the peak to declining stage of the disease (). With MMP13ap, we managed to visualize this later activity in vivo in CIA and show that it correlates quantitatively with disease progression. Another peptide substrate probe detecting MMP-13 activity was previously used in a rat model of osteoarthritis () and correlated qualitatively with the progression of cartilage degradation. These studies emphasize the potential of MMP13ap as a marker for quantifying cartilage degradation during arthritis.
Different MMPs are involved in various stages of both biologic and pathophysiologic processes (). With MMP12ap and MMP13ap, we have shown that it is possible to delineate these different activities in vivo. Protease-specific activatable probes that are detectable in vivo are important tools to monitor disease progression and the efficacy of protease-inhibitor therapeutics as they allow multiple measurements of net protease activity. Inhibitors of MMPs need to be highly selective and potent at low doses, as off-target effects and lowering of dosages have produced negative results in clinical trials (). MMP-specific activatable probes may be used during the testing of inhibitors, both to determine effective dosages and to study off-target effects. Within the limitations of fluorescence imaging technology, well-characterized and validated protease-specific activatable probes offer the potential to tailor treatment to particular phases of arthritic disease.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Nagase had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Lim, Meinjohanns, Gompels, Nuti, Rossello, Devel, Dive, Meldal, Nagase.
Acquisition of data. Lim, Meinjohanns, Bou-Gharios, Gompels, Meldal.
Analysis and interpretation of data. Lim, Bou-Gharios, Gompels, Dive, Meldal, Nagase.