|Home | About | Journals | Submit | Contact Us | Français|
We used a molecular probe activated by protease cleavage to image expression of matrix metalloproteinases (MMPs) in the heart following myocardial infarction.
We synthesized and characterized a near-infrared fluorescent (NIRF) probe that is activated by proteolytic cleavage by MMP2 and MMP9. The NIRF probe was injected into mice at various time-points up to 4 weeks after myocardial infarction induced by ligation of the left anterior descending artery. NIRF imaging of MMP activity increased in the infarct region with maximal expression at one to two weeks, persisting to four weeks. Zymography and real-time PCR analysis showed that MMP9 expression is increased at two to four days, while MMP2 expression is increased at one to two weeks. Dual label confocal microscopy showed colocalization of NIRF imaging with neutrophils on day 2, and flow cytometric analysis confirmed that NIRF signal is associated with leukocytes in the infarct zone.
This study demonstrates that the activity of MMPs in the myocardium may be imaged using specific activity-dependent molecular probes.
Myocardial infarction (MI) can lead to complex structural alterations in the heart, including dilation, contractile dysfunction and congestive heart failure1-4. Matrix metalloproteinases (MMPs) play important roles in the structural and functional sequelae of cardiac ischemia and infarction. In the heart, as in other tissues, MMPs participate in degradation of the extracellular matrix (ECM). In addition, MMPs have additional effects independent of ECM breakdown, including degradation of troponin I5, and modulation of platelet aggregation6,7 and vascular tone8,9.
The expression and activities of MMPs increase in the heart following MI10-13. Specifically, the gelatinases MMP2 and MMP9, and the tissue inhibitors of matrix metalloproteinases (TIMPs) have been implicated in the cardiac response to ischemia and infarction. Gene knockout of MMP214 or MMP915 decreases left ventricular remodeling and limits ventricular dilation after myocardial infarction. Loss of TIMP4 leads to enhanced MMP activity in the heart following ischemia16, and gene knockout of TIMP-1 exacerbates ventricular remodeling17. Pharmacologic blockade of MMP activity leads to decreased ventricular remodeling18-21.
Noninvasive methodology to optically image MMP activity would be of use to follow the temporal and spatial patterns of MMP activity after MI and in assessing treatment approaches. We hypothesized that we could monitor the myocardial remodeling process by imaging MMP activity. We used a novel, long-circulating, quenched NIRF probe that is activated by proteolytic cleavage22,23 by MMP2 and MMP9. Our studies indicate that specific molecular probes can be used to image MMP activity during left ventricular remodeling. This approach complements existing methods for following MMP levels and activity, such as zymography, immunohistochemistry, Western blotting, and scintigraphy24.
All procedures were approved by and performed in accordance with the IACUC of the Massachusetts General Hospital. Male C57BL6J mice, ranging in age from 8 to 12 weeks, weighing 25-30 g, underwent left anterior descending artery (LAD) ligation to induce MI25. Animals were anesthetized intraperitoneally with ketamine 0.065mg/g body weight, acepromazine 0.001 mg/g, and xylazine 0.013 mg/g. Animals were intubated and ventilated with a rodent respirator. After thoracotomy, the LAD was ligated with a 7-0 silk suture at a location 3-4 mm from the tip of the left atrium. Successful ligation of the LAD was verified by visual inspection of the left ventricular apex. The chest was closed with continuous 6-0 silk suture, and the skin was closed with 4-0 silk sutures. All mice were given water and food ad libitum.
The NIRF probe used in this study was synthesized using a previously published method26, except a peptide sequence recognized by MMP2 and MMP9 was used. This sequence, SGKGPRQITA, was identified by screening a phage library of random sequences using MMP9, and is cleaved between the Gln and Ile residues. It has a kcat/Km of 188,000 M−1s−1, which is 12 to 14 fold higher for MMP9 than for MMP-13 and MMP-727. This sequence was used with the linking residues GGPRQITAGK(Fitc)C to attach the fluorochrome Cy5.5 to a pegylated poly-L-lysine backbone.
Activation of the probe (0.2μM) was studied by incubating with 1 μg recombinant MMP9 or MMP2 (Oncogene) in 50 mM Tris HCl pH 7.5, 10 mM CaCl2, 100 mM NaCl and 0.005% Brij-35 at 37°C. The fluorescence signal changes were monitored at the specific excitation (675 nm) and emission (694 nm) maximum wavelengths of Cy5.5, using a fluorescence plate reader (SPECTRAmax Gemini, Molecular Devices, Sunnyvale, CA).
4 nmol NIRF probe was injected into the tail vein of mice 0 days, 2 days, 1 week, 2 weeks, 3 weeks and 4 weeks after myocardial infarction. This dose was chosen based on prior experience with other enzyme activatable probes in mice. The hearts were excised 48 hours later. NIRF imaging (2 minute acquisition) was performed using a 12-bit monochrome CCD camera (Kodak) equipped with an f/1.2 12.5-75 mm zoom lens and an emission long pass filter at 700 nm (Omega Optical). The NIRF signal was determined as mean signal intensities (SI) from either the infarct region or remote myocardium. Target-to-background ratios (TBR) were calculated as follows: TBR=SI(infarct area)/SI (remote area)28. All results are presented as mean ± SEM. Statistical analysis was conducted by using ANOVA, with values of P <0.05 being considered significant.
Hearts were excised and washed in cold PBS. The infarct region and remote myocardium were separated under a dissecting microscope. Tissues were snap frozen in liquid nitrogen and homogenized in 0.5-1.0 ml of ice cold lysis buffer. The Western blot lysis buffer contained 50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM PMSF, 5 ng/ml aprotinin, 5 ng/ml leupeptin, 1% Triton X-100, 1% sodium deoxycholate, and 1% SDS. The homogenates were centrifuged for 10 min at 4°C at 20,000g, and the supernatant was transferred to clean tubes. Protein concentration was determined by using a protein assay (Bio-Rad, CA) using bovine serum albumin as standards. The extracted samples were stored at −80°C until use.
Protein samples containing 25 μg protein were adjusted to 40 μl, mixed with an equal volume of 2× loading buffer (126 mM Tris-HCl, 20% glycerol, 4% SDS, 0.05% bromophenol blue, pH 6.8), and incubated for 10 min at room temperature prior to loading to 10% gelatin-ready zymogram gel (Bio-Rad, 161-1167). 10 μl Precision protein standards (Bio Rad, 161-0372) were used as molecular weight standards, and 0.5 ng recombinant human MMP9 protein (Oncogene, PF024) and 0.5 ng purified MMP2 (Chemicon, cc071) were loaded as positive controls. Gels were run at 80 volts for 2 hours. The gels were incubated in 2.5% Triton X-100 for renaturation for 30 min. They were then equilibrated in fresh developing buffer (50 mM Tris-HCl,0.2 mM NaCl, 5 mM CaCl2, 0.02% Brij 35) and incubated at 37 °C overnight. Gels were stained with 0.5% Coomassie Blue G (40% methanol, 10% acetic acid) for 2 hours and destained with methanol:acetic acid:water (5:1:4) for 30 min. Areas of protease activity appear as clear bands against a dark blue background where the protease has digested the gelatin substrate.
MMP9 and MMP2 mRNA levels were measured by quantitative real time PCR of cDNA29. Total RNA was extracted from infarct regions using Trizol (GIBCO-BRL) and reverse-transcribed to cDNA by using Superscript II RT (GIBCO-BRL) using oligo (dT) as primers. Real time PCR was performed using an ABI Prism 7000 Sequence Detection system using SYBR Green. Primers were designed using Primer Express software. The primers for beta actin were TGG AAT CCT GTG GCA TCC ATG AAA C (forward), and TAA AAC GCA GCT CAG TAA CAG TCC G (reverse). The primers for MMP2 were GAC ATA CAT CTT TGC AGG AGA CAA G (forward) and TCT GCG ATG AGC TTA GGG AAA (reverse). The primers for MMP9 were CCT GGA ACT CAC ACG ACA TCT TC (forward) and TGG AAA CTC ACA CGC CAG AA (reverse). For each primer set, a standard curve was generated between the threshold cycle number and cDNA concentration. MMP2 and MMP9 mRNA levels were expressed as a ratio normalized to beta actin. Melting curves of the PCR products and agarose gel electrophoresis confirmed the specificity of the reaction.
8 μm frozen sections were used for histology (hematoxylin and eosin staining), or immunohistochemistry for MMP9 and markers for macrophage and neutrophils. The following antibodies were used: (1) mouse anti-human MMP9 IgG (1:100, Chemicon, MAB 13415); (2) rat anti-mouse Ly-6G (Gr-1) IgG-FITC (1:50, PharMingen, 553126); (3) rat anti-mouse Mac3 IgG-FITC (1:50, PharMingen, 01784D); (4) donkey anti-mouse IgG-texas Red (1:50, Jackson, 715-075-150). The staining procedures have been described in detail elsewhere30. Confocal imaging was performed by confocal microscope (Leica TCS NT4D) with the use of a BP530/30 filter for FITC and a LP590 filter for Texas red. When the FITC and Texas Red signals colocalize, a yellow signal is seen.
Mice with infarcted hearts or control mice were injected with NIRF probe or vehicle control. 24 h following probe administration, animals were euthanized using CO2 asphyxiation and their hearts immediately excised and washed in Hank's Balanced salt solution (HBSS). Infarct and non-infarct heart tissues were separated. All tissues were then minced finely and digested in collagenase (Worthington Biochemical, Lakewood, NJ, 2 mg/ml (w/v) in HBSS) at 37°C for 45 min. The resulting cell suspensions were filtered through 100 um pore size filters, pelleted and resuspended in RPMI-1640 containing 5 % FCS. Cells were then stained with anti-mouse CD45-FITC conjugated antibody (BD Pharmingen, San Diego, CA). Each population was analyzed on a BDFACSCalibur flow cytometer for the presence of CD45 and the NIRF probe. A miminum of 10,000 cells per sample were analyzed.
The NIRF probe used in this study was similar to those used previously for imaging other proteases22,26,31. The probe contains multiple NIR fluorescent Cy5.5 dye molecules attached to the recognition peptide sequence which is positioned on a pegylated poly-L-lysine backbone. In the intact probe, the Cy5.5 molecules are in close proximity, and fluorescence resonance energy transfer results in efficient quenching. In the presence of MMP2 or MMP9, the peptide sequence is cleaved, resulting in liberation of the Cy5.5 dye molecules from the probe.
We tested the activation of the NIRF probe by MMP2 and MMP9 (Figure 1). At zero timepoint, there was only background fluorescence activity. Both MMP2 and MMP9 caused time-dependent increases in fluorescence, with roughly equivalent kinetics. Overall, either protease resulted in 200-fold signal amplification. Therefore, the NIRF probe is equally sensitive to MMP2 and MMP9.
To image MMP2 and MMP9 activity in the heart following MI, we performed LAD ligation in mice, resulting in an apical infarct. Figure 2 shows the appearance of the heart two days and one week later. The area of infarction is pale compared to surrounding tissue. At one week, the left ventricular apex showed thinning. The MMP2 and MMP9 sensitive probe was injected and NIRF imaging was performed. In the control animal, infarcted area had similar signal as normal tissue, with a target to background ratio (TBR) of 1.08±0.11. In animals receiving the probe, the TBR increased to 3.34±0.43 one week after MI, and areas of infarction were intensely bright (Figure 2).
We imaged MMP2 and MMP9 activity in different cohorts of animals to establish a time course over a four week period. The NIRF signal in the infarct zone rises and peaks at one to two weeks after myocardial infarction (Figure 3). It then decreases over time, but was still significantly elevated at 4 weeks compared to baseline values, or values from remote myocardium. Quantitation of the absolute NIRF signal, as well as the ratio of the signal between the infarct zone and a remote area, shows the same pattern.
To correlate the NIRF imaging results with expression of MMP2 and MMP9, we performed zymography and quantitative RT-PCR analysis. Zymography showed that MMP9 activity is increased in the infarct zone at two and four days post-MI, with levels dropping subsequently (Figure 4). Of interest, MMP9 activity is increased in the remote area as well at four days, though less than in the infarct zone. In the infarct region, MMP2 activity increases by one week and reaches a maximum at two and three weeks. MMP2 activity is present in the remote area at low, but detectable levels. Multiple molecular species are seen for MMP9 and MMP2.
We used real time PCR analysis of cDNA reverse transcribed from cardiac mRNA to quantitate MMP2 and MMP9 expression (Figure 5). In the infarct region, MMP9 mRNA levels were highest at two and four days post-MI, and decreased subsequently. MMP2 levels, in contrast, rose by four days, peaked at one to two weeks post-MI, and were still significantly increased at 4 weeks. Taken together, zymography and real-time PCR results were in agreement that MMP9 levels were increased at one and three days post-MI, and that MMP2 levels increased at one to four weeks.
Comparison of microscopic NIRF imaging with hematoxylin and eosin staining of an adjacent section shows that the fluorescence signal correlates with regions of leukocyte infiltration (Figure 6). Confocal microscopy with dual labeling for cell-specific markers and for MMP9 confirmed colocalization of MMP9 with a neutrophil marker at 1 day (Figure 7). To verify that MMP9 was leukocyte-derived in the infarcted myocardium, we made use of the NIRF probe to perform flow cytometric analysis on cells isolated from the heart. We used NIRF to detect the gelatinase probe, and immunoreactivity to the leukocyte common antigen CD45 to detect leukocytes. In the non-infarcted heart tissue, the majority of cells that display NIR fluorescence are CD45 negative. In the infarct region however, there is an increase in the number of CD45 positive cells that display NIR fluorescence (Figure 8). These results confirm our macroscopic observations and point to infiltrating leukocytes as the main source of the increase in MMP9 activity following myocardial infarction.
The MMPs are a family of Zn containing metalloproteinases that play important roles in ECM degradation32. MMPs have been divided into subgroups based on substrate specificity and domain structure33. MMP2 and MMP9 are of particular interest because they break down collagen, elastin, and basement membrane components. They share a common domain structure that includes type II fibronectin like domains, differentiating them from other MMPs34. Both MMP2 and MMP9 also have effects that appear to be independent of ECM breakdown. MMP2 degrades troponin I, a mechanism important to its effects on the heart following ischemia5. Other MMP targets affect vascular tone8,9 and platelet aggregation6.
Several lines of evidence point to the importance of MMP2 and MMP9 in cardiac remodeling after MI10-13, as well as in vascular remodeling following injury and during atherogenesis32,35. Gene knockout of either MMP214 or MMP915 results in attenuation of LV dilation following MI. Similarly, gene knockout of the MMP inhibitor TIMP-1 worsens ventricular dilation and remodeling after MI. Furthermore, pharmacologic blockade of MMP activity results in attenuation of ventricular dilation18-21.
Our results indicate that MMP9 and MMP2 activity show different time courses after MI. MMP9 increases by days 2 to 4, while MMP2 activity increases between 1 and 2 weeks. Other studies show similar time courses for MMP2 and MMP9 following MI36. Despite their shared substrate, MMP2 and MMP9 may play distinct roles following tissue injury, as shown in vascular remodeling32,35. In terms of cellular source, early after MI, NIRF activity colocalizes with areas of leukocytic infiltrate (Figure 6), and with neutrophils (Figure 7). Flow cytometry confirms that CD45-positive leukocytes account for the increase in NIR fluorescent cells in the heart following MI (Figure 8). These results indicate that infiltrating neutrophils are the likely cellular source of MMP9 activity early after MI.
Monitoring of MMP activity is complicated by overlapping substrate specificities of the MMPs, the fact that there are multiple species including proMMP zymogens, and multiple levels of regulation by TIMPs and other MMP inhibitors37. In addition, a recent report demonstrates that MMPs can be activated by S-glutathiolation in the presence of peroxynitrite38. Here, we show that an optically detectable probe that is activated by MMP2 and MMP9 can be used to image the activities of these proteases in the heart following MI. The use of such enzyme-activated probes would allow monitoring of the spatial and temporal patterns of MMP activity over time in the heart, and may facilitate the study of cardiac remodeling. They also provide a high degree of molecular specificity, as their fluorescence is based on activation by enzyme activity. However, the NIRF probe described here, like gelatin zymography, does not distinguish between MMP2 and MMP9.
In summary, we have shown that enzyme-activated NIRF probes can be used to image MMP activity in the myocardium following MI, providing both spatial and temporal information on activity. These probes may be useful to complement other techniques used to study MMPs, including zymography and scintigraphy24. Our results also indicate that these probes can be used to monitor enzyme activity on a microscopic level, and be coupled with flow cytometry. Similar NIRF probes have been coupled with optical tomography methods for in vivo imaging28,39.
This work was supported by PHS grants HL52818 to PLH and CA86355 to RW.