Critical Role of Mast Cell Chymase in Mouse Abdominal Aortic Aneurysm Formation

Human chymases, which are mast cell-restricted serine proteases, can activate matrix metalloproteinases (MMPs),¹ produce angiotensin II,² induce endothelial cell (EC) and smooth muscle cell (SMC) apoptosis,^3,4 and degrade constituents of high-density lipoprotein particles, thereby impairing their cholesterol efflux ability.⁵ Therefore, mast cell chymases may contribute to arterial remodeling in atherosclerosis and abdominal aortic aneurysm (AAA) formation. AAA involves extensive arterial extracellular matrix degradation, aortic cell apoptosis, and microvessel accumulation.⁶ To date, we lack a clinically useful soluble biomarker for AAA, and invasive repair is the only treatment for this life-threatening human disease.⁷ Although a direct link between chymase and AAA remains conjectural, extracts from human aneurysmal aortas have significantly higher mast cell chymase–associated angiotensin II–forming activity than those from control aortas,⁸ consistent with the observation of high numbers of chymase-positive mast cells in the adventitia and media of aneurysmal aortas.⁹ We recently established an important role of mast cells in AAA in 2 mouse preparations.¹⁰ Although the role of mast cell–derived chymases in the pathogenesis of AAA remains untested, mast cell–deficient Kit^W-sh/W-sh mice had greatly attenuated AAA formation after aortic elastase perfusion and periaortic CaCl₂ injury compared with wild-type (WT) control mice. Inhibition of mast cell chymase activities with a small-molecule inhibitor, NK3201, reduced elastase perfusion–induced AAA formation in dogs, hamsters, and mice.^11–13 Reduced AAA formation occurs with lower MMP-9 activity and mast cell numbers in the aortas. These data suggested that mast cell chymases contribute importantly to AAA formation, although detailed mechanisms remain unclear, and this inhibitor could have pertinent nonspecific effects.

Protease-deficient animals permit direct testing of the functions of individual enzymes in pathological processes, including AAA.^14,15 Unlike human chymase, the product of a solitary gene belonging to the α-chymase family, mice have a chymase family consisting of mouse mast cell proteases (mMCP)-1, -2, -4, -5, and -9. The α-chymases mMCP-1 and -2 are expressed mainly in mucosal mast cells, whereas the α-chymase mMCP-9 is expressed mainly in the uterus. In contrast, the connective tissue mast cells in most sites express the β-chymase mMCP-4 and the α-chymase mMCP-5. mMCP-4 appears to be the predominant mast cell chymase within mouse connective tissues.¹⁶ Therefore, mMCP-4 may be the most relevant human chymase ortholog in arterial remodeling. This study used mMCP-4–deficient (Mcpt4^−/−) and mMCP-5–deficient (Mcpt5^−/−) mice to test the hypothesis that these connective tissue chymases impair AAA formation directly in mice with AAA induced by aortic elastase perfusion.

Methods

Human AAA Lesion Sections and Serum Samples

Paraffin-embedded human aortic sections were prepared from 10 AAA donors (5 females and 5 males; mean age, 78.80±2.05 years) and 10 non-AAA heart transplant patients (5 females and 5 males; mean age, 41.90±4.19 years) without detectable vascular diseases from the Department of Surgery, Washington University, St Louis, Mo. These sections helped to detect chymase expression with the use of mouse anti-human chymase monoclonal antibody (Abcam). Human aortic tissue extracts were prepared from 3 female AAA patients and 3 female heart transplant donors with no detectable vascular disease from the Department of Medicine, Brigham and Women’s Hospital, Boston, Mass. Tissue lysates were used for immunoblot analysis (30 μg per lane) with the same antibody. The same blot was reprobed for β-actin to affirm equal protein loading. Separate human protocols were preapproved by the Human Investigation Review Committees at Washington University and at Brigham and Women’s Hospital.

To establish the correlation of AAA expansion rates to serum chymase levels, we developed a human chymase enzyme-linked immunosorbent assay system. Briefly, mouse anti-human chymase antibody (AbD, Serotec) was used to coat a 96-well plate. Serum samples were diluted (1:100) and incubated on a precoated plate for 2 hours at room temperature. Biotinylated mouse anti-human chymase monoclonal antibody (Millipore) was used as detecting antibody. Detailed antibody information is listed in Table I in the online-only Data Supplement.

Serum samples from 103 male AAA patients were identified by a previously described population-based screening of 65- to 73-year-old men at the Department of Vascular Surgery, Viborg Hospital, Viborg, Denmark.¹⁷ A total of 4404 men were invited to participate in this follow-up study. Of 3344 participants, 141 (4.2%) had AAA, defined as an infrarenal aortic diameter of ≥30 mm. Of those with AAA, 19 had diameters >50 mm who were referred for surgery, and 10 were lost to follow-up. A total of 112 were followed for >1 year (on average, for 2.9 years). Of 112 cases, 103 had serum chymase measured blindly and data presented. All participants gave informed consent. Detailed patient information was reported previously.¹⁷

Mouse AAA Model and Lesion Characterization

C57BL/6 WT mice were purchased from The Jackson Laboratory (Bar Harbor, Me). Mcpt4^−/− and Mcpt5^−/− C57BL/6 mice have been backcrossed to the same background for >10 and 4 generations, respectively.^18,19 Ten-week-old mice from each type were used for aortic elastase perfusion–induced AAA.¹⁴ Aneurysmal lesions were collected 7, 14, and 56 days after elastase perfusion. Mouse aortic diameters were measured before and after elastase perfusion and at the end of each time point. Aortic diameter expansion ≥100% of that before perfusion defined AAA, according to Pyo et al.¹⁴ Each mouse aorta was isolated for both frozen section preparation and tissue protein extraction in a pH 5.5 buffer.¹⁷ Frozen sections were used for immunostaining for macrophages (Mac-3), SMCs (α-actin), T cells (CD3), apoptotic cells (terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling [TUNEL]), microvessels (CD31), chemokine (monocyte chemotactic protein-1 [MCP-1]), mast cells (c-Kit, CD117), and elastin degradation (Verhoeff–van Gieson) (Table I in the online-only Data Supplement). SMC content and elastin fragmentation were graded as described previously.¹⁷ T cells, apoptotic cells, mast cells, MCP-1–positive cells, and microvessel numbers were counted blindly. Macrophage-positive areas were measured with the use of computer-assisted image analysis software (Image-Pro Plus; Media Cybernetics, Bethesda, Md).

Bone Marrow–Derived Mast Cell Preparation and KitW-sh/W-sh Reconstitution

Bone marrow–derived mast cells (BMMCs) were prepared and verified as described previously.²⁰ For Kit^W-sh/W-sh mice reconstitution, recipient mice at 5 weeks of age were given 1×10⁷ of BMMCs from WT or Mcpt4^−/− mice via the tail vein. Five weeks after BMMC reconstitution, mice were introduced to the AAA model. Mouse abdominal aortas were harvested 14 and 56 days after elastase perfusion.

BMMC In Vitro Activity Assays

An aortic ring angiogenesis assay was performed by incubating Matrigel-embedded WT mouse aortic rings in a 96-well plate with live BMMCs from WT and Mcpt4^−/− mice (3×10⁵ cells per 150 μL RPMI-1640 with 10% fetal bovine serum per well) for 7 to 10 days. EC growth areas were determined with the use of Image-Pro Plus software and presented as square millimeters. Vascular endothelial growth factor (VEGF) (10 ng/mL; PeproTech) was used as a positive control.

Real-time polymerase chain reaction (RT-PCR) was used to determine protease mRNA levels in BMMCs. Total cellular RNA was extracted from BMMCs with the use of TRIzol reagent (GIBCO) and treated with RNase-free DNase (Ambion) to remove genomic DNA contaminants. Equal amounts of RNA were reverse-transcribed, and quantitative PCR was assessed in a single-color RT-RCR detection system (Stratagene). The level of each protease transcript was normalized to that of the β-actin transcript.

BMMC and aortic tissue extract cysteinyl cathepsin activities were examined with biotin-conjugated JPM, an affinity probe that labels active cathepsins specifically and irreversibly.¹⁷ Briefly, BMMC and pulverized aortic tissues were lysed into a pH 5.5 buffer. Five micrograms of protein from each sample was incubated with 12 mmol/L dithiothreitol and 1 μL of biotin-conjugated JPM in 30 μL of a pH 5.5 buffer for 1 hour at 37°C. Protein samples were then separated on a 12% SDS-PAGE followed by immunoblot detection with horseradish peroxidase–conjugated avidin (Table I in the online-only Data Supplement). BMMC lysate MMP activity was detected with a gelatin gel zymogram, essentially the same as reported previously.²¹

BMMC activity in promoting vascular SMC apoptosis was performed with the use of primary cultured mouse aortic SMC on an 8-well chamber slide.¹⁰ Confluent SMCs were stimulated to apoptosis overnight with 80 μmol/L pyrrolidinedithiocarbamate with and without 300 μL of degranulated BMMC supernatant in Dulbecco’s modified Eagle’s medium (1×10⁶ BMMC supernatant/mL) from WT or Mcpt4^−/− mice as described.²² Apoptotic cells were detected with an In Situ Cell Death Detection Kit according to the instructions (Roche Diagnostics Co).

BMMC elastase activities were determined by mixing 100 μg of BMMC extract with 20 μg of fluorogenic elastin (DQ-elastin, Molecular Probes, Carlsbad, Calif) in 100 μL of a pH 5.5 buffer on a 96-well plate for 5 hours. The plate was read at excitation/emission=505/515 nm, and data were presented as relative fluorescence units.

Statistical Analysis

The annual growth rate of human AAA during the first year after serum sampling was calculated as the difference between maximal anterior-posterior AAA diameter divided by the observation time. The crude correlation between serum chymase, AAA size, and growth rate was calculated with Pearson correlation test. Multivariate linear regression analysis was performed to adjust for the subjects’ use of glucocorticoids, smoking, low-dose aspirin, angiotensin-converting enzyme inhibitor, β-adrenergic blocker, β-adrenergic agonist, lowest ankle-brachial blood pressure index, diastolic blood pressure, and age.

Because of small sample size and abnormal data distribution, a Mann–Whitney U test was used for data from mouse experiments throughout the study. P values <0.05 were considered statistically significant.

Results

Increased Chymase Expression in Human AAA Lesions

Human AAA lesions had many more chymase-immunoreactive mast cells than did healthy aortas. Chymase-positive mast cells localized throughout the AAA lesions from media to adventitia (Figure 1A and 1B). In contrast, the media and adventitia of human aortas without AAA had many fewer chymase-positive mast cells (Figure 1C). Consistent with this observation, immunoblot analysis of aortic tissue extracts showed 11-fold higher levels of chymase proteins by densitometry in AAA than in normal aortas after the protein loading for β-actin signals was normalized (ImageJ) (Figure 1D).

• Download figure
• Download PowerPoint

Correlation of Serum Chymase Level With AAA Growth Rate

In this serological prospective study of 103 AAA patients who had serum chymase measured, Pearson correlation test did not reveal significant associations between chymase and initial AAA size (r=0.02, P=0.833). However, serum chymase levels correlated modestly with AAA growth rate (r=0.19, P=0.06). Because of the various medications used by these patients, we performed a multivariate linear regression analysis by adjusting for potentially known confounders of AAA, including smoking, low-dose aspirin, use of angiotensin-converting enzyme inhibitors, β-adrenergic blocking agents, β-adrenergic agonists, or corticosteroids, ankle-brachial index, diastolic blood pressure, and age. Among these confounders, serum chymase and glucocorticoid use correlated strongly and significantly with AAA growth rate after the adjustment (P=0.009) (Table). Although we do not know the cause, glucocorticoid users had higher AAA growth rate (3.54±0.49 versus 1.97±0.19 mm/y; P=0.01) but lower serum chymase levels (10.77±0.31 versus 12.34±0.24 ng/mL; P<0.04) than nonglucocorticoid users. Therefore, chymase expression alone may not explain the glucocorticoid effects on AAA growth. Other unrecognized mechanisms may play more important roles. After those receiving glucocorticoid treatments were excluded (n=10), the modest correlation between serum chymase and AAA growth rate became significant (r=0.26, P=0.01), although we did not see any correlation among the glucocorticoid users (r=0.07, P=0.84) (Figure 1E).

Reduced AAA Formation in mMCP-4–Deficient Mice

Increased chymase-positive mast cells in human AAA lesions and significant association of serum chymase levels with AAA growth rate suggested the involvement of chymases in AAA development. To test this hypothesis directly, we performed elastase aortic perfusion to induce AAA in Mcpt4^−/− and Mcpt5^−/− mice.¹⁴ Analysis at 7 and 14 days after perfusion did not show significant differences in aortic expansion between WT and Mcpt5^−/− mice (data not shown). In contrast, whereas all WT mice developed AAA 14 days after elastase perfusion (100% incidence), none of the Mcpt4^−/− mice did (0% incidence) (P<0.0001). Analysis 56 days after perfusion still showed significant attenuation of AAA expansion (P=0.0003) and lower incidences (69% versus 100%) in Mcpt4^−/− mice compared with WT mice (Figure 2A). Inflammatory cell accumulation, including Mac-3⁺ macrophages (Figure 2B) and CD3⁺ T cells (Figure 2C), also diminished in Mcpt4^−/− mice relative to WT control mice at the 14-day or 56-day time point. In contrast, aortas from Mcpt4^−/− mice had more SMCs than those from WT mice 7 days after perfusion (Figure 2D), although such differences diminished at later time points.

• Download figure
• Download PowerPoint

Apoptosis, angiogenesis, and elastin loss characterize human AAA.⁶ To examine whether reduced AAA in Mcpt4^−/− mice also impaired these variables, we performed TUNEL (apoptosis), CD31 (angiogenesis), and Verhoeff–van Gieson (elastin) staining with frozen aortic sections prepared from WT and Mcpt4^−/− mice from different time points. As we anticipated, the numbers of both apoptotic cells and CD31⁺ microvessels as well as the degree of elastin fragmentation fell significantly in AAA lesions from Mcpt4^−/− mice compared with those from WT control mice at most of the time points examined, although not at all time points (Figure 2E through 2G), suggesting that multiple variables contributed to reduced AAA in Mcpt4^−/− mice and that chymase expression affected these variables differently at assorted time points. Reduced AAA in Mcpt4^−/− mice did not result from altered accumulation of mast cells in AAA lesions. Lesion MCP-1⁺ or CD117⁺ cell content did not differ between the 2 groups (data not shown).

Mast Cell Reconstitution and AAA Formation in Mast Cell–Deficient Mice

Reduced AAA in Mcpt4^−/− mice might not result solely from the lack of mMCP-4 because aside from mast cells, mMCP-4 deficiency may affect other cell types, which could contribute indirectly to AAA formation. To examine this hypothesis, we reconstituted mast cell–deficient Kit^W-sh/W-sh mice with BMMCs from WT or Mcpt4^−/− mice. Consistent with our earlier observation, Kit^W-sh/W-sh mice failed to develop AAA,¹⁰ and provision of WT BMMCs restored AAA expansion at both the 14- and 56-day time points. In contrast, those that received BMMCs from Mcpt4^−/− mice remained protected from aortic expansion at both time points (Figure 3A).

• Download figure
• Download PowerPoint

In addition to measuring aortic diameters for those BMMC-reconstituted mice, we also determined their lesion macrophage and T cell content, CD31⁺ microvessel number, total apoptotic cell number, and elastin fragmentation. WT BMMC-reconstituted Kit^W-sh/W-sh mice restored at least partially these variables at both time points. In contrast, most variables did not differ significantly between Kit^W-sh/W-sh mice and those that received Mcpt4^−/− BMMCs at either time point (Figure 3B through 3F), supporting the hypothesis that reduced AAA in Mcpt4^−/− mice resulted from the absence of mMCP-4 from mast cells.

Chymase mMCP-4 Stimulates Microvascularization and Vascular Cell Apoptosis

Microvascularization and cell death occur in human AAA. Impaired angiogenesis and apoptosis in Mcpt4^−/− mice (Figure 2F and 2G) or in Kit^W-sh/W-sh mice that received Mcpt4^−/− BMMCs (Figure 3D and 3F) could result from reduced AAA formation in these mice. To examine whether mMCP-4 participated directly in angiogenesis and vascular cell apoptosis, we performed aortic ring microvessel outgrowth and SMC apoptosis assays with the use of BMMCs from WT and Mcpt4^−/− mice. Relative to WT BMMCs, those lacking mMCP-4 showed greatly reduced activity in promoting aortic ring microvessel outgrowth (Figure 4A). Earlier studies suggested that mast cells contribute to microvessel growth via VEGF²³ but less so by releasing inflammatory cytokines interleukin-6 (IL-6), interferon-γ, and tumor necrosis factor-α.¹⁰ To test whether mMCP-4 deficiency affects mast cell VEGF expression, we performed a VEGF enzyme-linked immunosorbent assay on supernatants from degranulated mast cells and lysates from WT and Mcpt4^−/− BMMCs and found no significant differences (data not shown). We have previously shown that WT BMMCs promote vascular SMC apoptosis.¹⁰Mcpt4^−/− BMMCs conferred significant protection to SMCs from pyrrolidinedithiocarbamate-induced apoptosis (Figure 4B), suggesting a direct role of mMCP-4 in both microvascularization and apoptosis during the development of experimental mouse AAA.

• Download figure
• Download PowerPoint

Deficiency of mMCP-4 Reduces Cysteine Protease Cathepsin Expression in Mast Cells

Angiogenesis and apoptosis both require protease activities. We have shown that the cysteine protease cathepsin S can modulate microvessel growth.^24,25 MMPs promote angiogenesis by triggering the release of VEGF.²⁶ Cysteinyl cathepsins also participate in apoptosis by cleaving the Bcl-2 family member Bid followed by mitochondria cytochrome c release.²⁷ Likewise, MMPs participate in cell death although via different mechanisms.²⁸ Considering reduced angiogenesis and apoptosis in Mcpt4^−/− mice (Figure 2C and 2F) or in Mcpt4^−/− BMMC-reconstituted Kit^W-sh/W-sh mice (Figure 3D and 3F), we hypothesized that absence of mMCP-4 affects cathepsin and/or MMP expression. Mcpt4^−/− BMMCs had significantly lower mRNA encoding all cathepsins tested, including B, S, K, and L, compared with those in WT BMMCs, as shown by RT-PCR. Although both MMP-2 and -9 mRNAs were also higher in WT BMMCs than in Mcpt4^−/− BMMCs, their RNA levels remained low in WT BMMCs (Figure 4C). MMPs in Mcpt4^−/− BMMCs were undetectable by gelatin zymography (data not shown). To document further decreased cathepsin activities in Mcpt4^−/− BMMCs, we performed cysteine protease active site labeling with biotin-JPM. Mcpt4^−/− BMMCs showed reduced overall cathepsin activities relative to WT BMMCs (Figure 4D). Interestingly, we detected high levels of cathepsin L mRNA but low levels of its active enzyme in WT BMMCs (Figure 4C and 4D), suggesting substantial regulation at the translation and/or activation levels. Assay of elastin degradation in vitro supported this observation. An equal amount of cell lysate from Mcpt4^−/− BMMCs degraded significantly less elastin than those from WT BMMCs under conditions optimized for cysteine protease cathepsins²⁹ (Figure 4E), suggesting that absence of mMCP-4 not only affects mast cell cathepsin mRNAs but also their activities.

mMCP-4 Induces Aortic SMC Cathepsin Activities

Mast cells induce vascular cell cathepsin expression via inflammatory cytokines, and these cathepsins may in turn participate in aortic wall remodeling.²⁰ We tested whether Mcpt4^−/− BMMCs demonstrate impaired ability to induce vascular cell cathepsin expression. Culture of aortic SMCs with BMMCs showed that WT BMMCs induced the activities of all major cysteinyl cathepsins to a much greater extent than Mcpt4^−/− BMMCs. Mcpt4^−/− BMMC-treated SMCs had reduced cathepsin B activity and negligible cathepsin S/K and L/C activities compared with SMCs treated with WT BMMCs (Figure 4F). Thus, the absence of mMCP-4 in mast cells not only altered mast cell cathepsin expression and activities but also decreased the ability of mast cells to induce cathepsin activity in vascular SMCs by an unknown mechanism.

Deficiency of mMCP-4 Reduces Aortic Wall Cathepsin Activities

Reduced cathepsin activities in Mcpt4^−/− BMMCs and in Mcpt4^−/− BMMC-treated SMCs suggest that reduced development of AAA in Mcpt4^−/− mice resulted at least in part from reduced cathepsin activities in the vessel wall. We performed cathepsin active site labeling by incubating biotin-JPM with aortic tissue extracts from WT and Mcpt4^−/− mice. Consistent with our hypothesis, aortic tissues from Mcpt4^−/− mice contained much less cathepsin activity, especially cathepsins S/K, than those of WT mice (Figure 5A). In situ elastin zymography yielded similar results. Under conditions optimized for cathepsin activities (pH 5.5 with ethylenediaminetetraacetic acid),²⁹ frozen aortic sections from WT mice (Figure 5B) had greater activity in digesting fluorogenic elastin than those from Mcpt4^−/− mice (Figure 5C).

• Download figure
• Download PowerPoint

Discussion

We have previously shown that mast cells play an important role in mouse AAA formation.¹⁰ These cells produce cytokines, chemokines, and proteases, all of which may participate in the pathogenesis of AAA.^7–9 Like macrophages, mast cells produce large quantities of cathepsins B, S, K, L, and C.³⁰ In contrast with macrophages, however, mast cells express specialized proteases, including the serine protease chymases. These proteases can activate MMPs and process angiotensinogen to produce angiotensin II, mediators proven critical in mouse AAA formation.^14,31 Human AAA lesions contain chymases and their substrates; whether these proteases participate causally in the pathogenesis remains uncertain. This study supports the hypothesis that mast cell chymase participates directly in the pathogenesis of AAA formation by assisting microvessel growth, vascular cell apoptosis, and cysteinyl cathepsin expression and activation. Yet these findings raise unanswered questions.

Preoperative glucocorticoid therapy has been associated with benefit in isolated cases of inflammatory aneurysms.³² Such treatment in patients with AAA increases IL-10 (a generally anti-inflammatory cytokine) and decreases IL-6 (a proinflammatory cytokine) and levels of the inflammatory marker C-reactive protein in serum.^33,34 In contrast to these beneficial effects, our data showed that patients who received glucocorticoid therapy had a nearly doubled rate of AAA growth (mean±SEM, 3.54±0.49 versus 1.97±0.19 mm/y; P=0.01) compared with patients who did not receive the therapy, although this treatment also significantly reduced the serum chymase levels (mean±SEM, 10.77±0.31 versus 12.34±0.24 ng/mL; P<0.04). Although this population included only 10 patients receiving glucocorticoids, the data indicate the potential of glucocorticoids as a strong confounder in the analysis of chymase activity and AAA progression and therefore the multivariate analysis included these patients. The positive association between AAA growth and use of glucocorticoids could be confounded by indication. For instance, patients often use glucocorticoids because they have severe pulmonary obstructive disease, which independently is associated with increased aortic expansion rate and rupture risk.³⁵ Alternatively, glucocorticoids can inhibit vascular SMC proliferation³⁶ and extracellular matrix protein synthesis.³⁷ Thus, in glucocorticoid-treated AAA, an unfavorable balance between extracellular matrix degradation versus synthesis may prevail despite the steroid-induced muting of aspects of inflammation, a hypothesis that requires further investigation.

We found that mast cell chymase expression affected the expression of other proteases. Reduced cysteinyl cathepsin expression in Mcpt4^−/− BMMCs leads to the hypothesis that the mechanism for the reduced AAA in Mcpt4^−/− mice might result not only from the absence of mMCP-4 but also reduced mast cell cathepsin expression. Although mMCP-4 can activate procollagenase, mMCP-4 itself is not a collagenase or an elastase.³⁸ Chymase mMCP-5 has elastinolytic activity, but lack of mMCP-5 did not impair AAA formation (data not shown). Therefore, neither mMCP-4 nor mMCP-5 alone likely contributes to the massive aortic wall elastinolysis during AAA formation. In contrast, cathepsins S, K, L, and C are potent elastases and collagenases, and these secreted cysteine proteases contributed to AAA formation¹⁵ (G. Shi, DSc, unpublished data, 2009). Absence of mMCP-4 reduced an array of elastinolytic and collagenolytic cathepsins in mast cells, which may account in part for reduced elastinolysis in Mcpt4^−/− mice (Figures 2E and 3E). Besides these cysteine proteases, mMCP-4 deficiency also affected the expression of cathepsin G, a serine protease present in human AAA capable of elastin degradation and angiotensin II generation.^39–41 BMMCs from Mcpt4^−/− mice had significantly lower levels of cathepsin G transcripts than those in WT BMMCs after normalizing to β-actin transcripts (mean±SEM, 0.013±0.0005 versus 0.0006±0.0003; P<0.00005), as shown by RT-PCR. Thus, reduced mast cell cathepsin G expression may also contribute to the observed attenuation in AAA formation in Mcpt4^−/− mice.

Mast cell chymases promote angiogenesis⁴² and apoptosis.⁴³ However, these in vitro studies employed purified chymases in isolated cells. Direct involvement of chymase in these processes in vivo remains unknown. This study showed results similar to those of prior studies of angiogenesis and apoptosis both in vitro (Figure 4A and 4B) and in vivo (Figure 2F and 2G). Reduced cysteinyl cathepsin expression and activities in Mcpt4^−/− BMMCs (Figure 4C through 4E) and impaired ability of Mcpt4^−/− BMMCs to induce SMC cathepsin expression raised the possibility that mast cell chymases participate in arterial remodeling in part by regulating cathepsin activities, a hypothesis that remains to be tested.

One important finding that we currently cannot explain is the impact of mMCP-4 on cathepsin and MMP expression. Our earlier observations from cathepsin S–deficient ECs indicated that lack of cathepsin S did not change the expression or activities of other cathepsins, the cathepsin inhibitor cystatin C, MMP, or tissue inhibitors of MMPs.²⁴ Chymase might participate in cathepsin activation, as it does in the case of pro-MMPs,¹ a hypothesis consistent with the observations of decreased cathepsin activities in the absence of mMCP-4 in both interacting with biotin-conjugated JPM (Figure 4D) and in degrading matrix elastin (Figure 4E). However, other explanations may apply. Mast cells lacking mMCP-4 contained greatly less cathepsin and MMP mRNA than did WT mast cells, suggesting that mMCP-4 regulates transcription and/or RNA-stability factors that control cathepsin expression. A protease that regulates the transcription of other proteases has not been reported, but an indirect pathway is possible. For instance, mast cell tryptase stimulates IL-8 and IL-lβ expression by ECs.⁴⁴ Chymase may act in a similar manner. The resulting inflammatory cytokines can exhibit autocrine and paracrine effects on protease expression.

Although not shown in this study, we found significantly lower transmigration of BMMCs from Mcpt4^−/− mice through collagen I–precoated filters compared with BMMCs from WT mice, suggesting impaired mast cell recruitment or accumulation to AAA lesions. However, the AAA of Mcpt4^−/− and WT mice had similar MCP-1 levels and CD117⁺ mast cell numbers (data not shown). These data suggest that mast cell transmigration in vitro requires protease activities but that other factors play more important roles in mast cell accumulation during AAA development, a hypothesis that merits further study. In conclusion, increased mast cell chymase expression in human or mouse AAA lesions directly promotes the pathological progression in part by regulating pertinent protease expression and activities. Inhibition of chymase activities limited experimental AAA formation and may assist in attenuating the progression of this irreversible human aortic disease.

Footnotes