The balance between matrix metalloproteinases (MMPs) and their natural inhibitors, the tissue inhibitors of metalloproteinases (TIMPs), plays a critical role in cardiac remodeling. Although a number of studies have characterized the pathophysiological role of MMPs in the heart, very little is known with respect to the role of TIMPs in the heart. To delineate the role of TIMPs in the heart we examined the effects of adenovirus-mediated overexpression of TIMP-1, -2, -3, and -4 in cardiac fibroblasts. Infection of cardiac fibroblasts with adenoviral constructs containing human recombinant TIMP (AdTIMP-1, -2, -3, and -4) provoked a significant (P < 0.0001) 1.3-fold in increase in bromodeoxyuridine (BrdU) incorporation. Similarly, treatment of cardiac fibroblasts with AdTIMP-1-, -2-, -3-, and -4-conditioned medium led to a 1.2-fold increase in BrdU incorporation (P < 0.0001) that was abolished by pretreatment with anti-TIMP-1, -2, -3, and -4 antibodies. The effects of TIMPs were not mimicked by treating the cells with RS-130830, a broad-based MMP inhibitor, suggesting that the effects of TIMPs were independent of their ability to inhibit MMPs. Infection with AdTIMP-1, -2, -3, and -4 led to a significant increase in α-smooth muscle actin staining, consistent with TIMP-induced phenotypic differentiation into myofibroblasts. Finally, infection with AdTIMP-2 resulted in a significant increase in collagen synthesis, whereas infection with AdTIMP-3 resulted in a significant increase in fibroblast apoptosis. TIMPs exert overlapping as well as diverse effects on isolated cardiac fibroblasts. The observation that TIMPs stimulate fibroblast proliferation as well as phenotypic differentiation into myofibroblasts suggests that TIMPs may play an important role in tissue repair in the heart that extends beyond their traditional role as MMP inhibitors.
- matrix metalloproteinase
- phenotypic differentiation
previous studies from this (31) and other laboratories (13, 21, 32) have suggested that the interaction between matrix metalloproteinases (MMPs) and tissue inhibitors of metalloproteinases (TIMPs) plays a critical role in the process of cardiac remodeling. Although an extensive amount of work has been performed in terms of understanding and characterizing the pathophysiological role of MMPs in the heart, far less is known about the role of TIMPs in the heart. The TIMP family presently consists of four distinct members, known as TIMP-1, -2, -3, and -4, each of which is constitutively expressed in the heart by fibroblasts as well as myocytes (14, 20, 36). TIMPs-1, -2, -3, and -4 are secreted proteins that act as the natural inhibitors of active forms of all MMPs, although the efficiency of MMP inhibition varies among the different members. Although the canonical role that has been ascribed to TIMPs-1, -2, -3, and -4 in the heart is that of neutralizing active MMPs, there is increasing evidence that TIMPs may exert “nontraditional” effects that are independent of their ability to inhibit MMPs. For example, in noncardiac cell types TIMPs have been shown to both stimulate and inhibit cell growth, as well as to promote cell survival or programmed cell death (4). Accordingly, to better understand the role of TIMPs in the heart, we have used adenovirus-mediated overexpression of TIMP-1, -2, -3, and -4 in cardiac fibroblasts to delineate the role of TIMPs in this cardiac cell type. Here we report that TIMP-1, -2, -3, and -4 share several overlapping functions in cardiac fibroblasts, as well as exhibiting several divergent effects that are unique to TIMP-2 and -3.
Preparation of cardiac fibroblasts.
All animal procedures were performed in accordance with the “Guiding Principles for Research Involving Animals and Human Beings” of the American Physiological Society. Cultured fibroblasts were prepared as previously described from hearts from 4- to 8-wk-old C57BL/6 mice (37). Because the phenotype of fibroblasts can be influenced by growth conditions such as cell passage and cell density (3, 25), only cells from passages 4 and 5 were used for the present studies. Fibroblasts were seeded at a concentration of 4.2 × 103/cm2 in DMEM containing 10% FBS and 1% penicillin-streptomycin and then synchronized in serum-free DMEM containing 1% penicillin-streptomycin 24 h before treatment.
Characterization of adenoviral TIMP-1–4 construct-infected cardiac fibroblasts.
The adenoviral constructs containing human recombinant (rh)TIMP-1, -2, and -3 (AdTIMP-1, -2, and -3) and Rad66 (AdNull), a control adenovirus, were described previously (5, 6). Our laboratory previously showed (5, 6) high levels of secretion of each AdTIMP by reverse zymography and Western blotting. A recombinant adenovirus expressing TIMP-4 (AdTIMP-4) was constructed by subcloning the human TIMP-4 open reading frame from pCI-TIMP-4 (a gift from Dr. Y. Eric Shi, Albert Einstein College of Medicine, New Hyde Park, NY; Ref. 23) into the pAdEasy-1 cassette (a gift from Dr. Bert Vogelstein, National Institutes of Health, Bethesda, MD; Ref. 16) with PCR. Adenoviral recombinants were propagated in 293 cells and titered as previously described (16). An AdNull construct with green fluorescent protein was generated in an identical manner and used as the appropriate control adenovirus for AdTIMP-4.
To determine the optimal multiplicity of infection (MOI) for TIMP-1, -2, -3, and -4 expression, cardiac fibroblasts were infected for 48 h with AdTIMP-1, -2, -3, or -4 at an MOI of 0, 100, 1,000, and 2,500. The cells were harvested, and TIMP-1, -2, -3, and -4 expression was determined by Western blot analysis of cell lysates (37) using rabbit anti-human polyclonal antibody (1:1,000 dilution) for TIMP-1, -2, and -3 (Chemicon International, Temecula, CA) and a goat anti-human polyclonal antibody (1:1,000 dilution) for TIMP-4 (Santa Cruz Biotechnology, Santa Cruz, CA). To analyze the time course of TIMP-1, -2-, 3, and -4 secretion, cardiac fibroblasts were infected for 24, 48, or 72 h with AdTIMP-1, -2, -3, or -4 in serum-free medium, after which the medium was collected and concentrated with Centricon YM-10 filters (Millipore, Bedford, MA). TIMP-1, -2, -3, and -4 protein expression was evaluated with Western blot analysis as described above.
In preliminary studies we treated cultured cardiac fibroblasts with commercially available TIMP-1, -2, and -4, to confirm the effects of AdTIMP-1–4 on cardiac fibroblasts. However, these preliminary studies showed that whereas the rhTIMPs inhibited MMP activity, they had no discernable effect on fibroblast function. Therefore, we used AdTIMP-conditioned medium, as described below. Although the reasons for the lack of biological effect on rhTIMP-1, -2, and -4 on fibroblast function are not known, they may relate to the lack of glycosylation and/or protein folding of the rhTIMPs.
Effect of AdTIMP infection on cardiac fibroblast proliferation.
Cells were seeded in 96-well plastic plates at a final volume of 100 μl/well. Cell cultures were infected for 24, 48, or 72 h with AdTIMP-1, -2, -3, -4, or AdNull or treated with 25 ng/ml basic FGF, which was used as a positive control. Cardiac fibroblast proliferation was determined with a colorimetric cell proliferation ELISA bromodeoxyuridine (BrdU) kit (Roche, Indianapolis, IN) as described previously (37). In preliminary control experiments we determined that there was a linear correlation between BrdU incorporation and cell number obtained by direct manual counting of cells (data not shown).
Effect of AdTIMP-conditioned medium on cardiac fibroblast proliferation.
To generate AdTIMP-1-, -2-, -3-, and -4-conditioned media, cardiac fibroblasts were seeded onto six-well plastic plates and infected with AdNull or AdTIMP-1, -2, -3, or -4 in DMEM supplemented with 10% FBS, 1% penicillin streptomycin, and 1% l-glutamine. Eighteen hours after adenoviral infection, the fibroblast cultures were washed three times in PBS and cultured for an additional 72 h in 2 ml of serum-free DMEM to generate fibroblast-conditioned medium. To determine the specificity of the effects of conditioned media, an aliquot of AdTIMP-1-, -2-, -3-, or -4-conditioned medium was neutralized with 2 mg/ml anti-TIMP-1, -2, -3, or -4 polyclonal antibody. The TIMP-antibody complexes were then removed by adding 10 μl of protein A agarose beads (Santa Cruz Biotechnology) to the conditioned medium, followed by centrifugation at 3,500 rpm for 5 min. The resultant supernatant was collected and stored at −80°C. To determine the effects of AdTIMP-infected cultured medium, fibroblast cultures were prepared as described above in 96-well plates and then stimulated for 24 h with TIMP-1-, -2-, -3-, and -4-conditioned media. Fibroblast proliferation was determined by examining BrdU incorporation, as described above.
Effect of AdTIMP infection and AdTIMP-conditioned medium on cardiac fibroblast apoptosis.
Cardiac fibroblasts were seeded onto glass slides (untreated) and allowed to attach overnight. Cells were then treated with AdTIMP-1, -2, -3, -4, AdNull, or diluent for 72 h, as described above. In preliminary control experiments we determined that the prevalence of cell apoptosis was similar in AdNull-infected cardiac fibroblasts and diluent-treated controls. Parallel experiments were also conducted with AdTIMP-1-, -2-, -3-, or -4-conditioned medium for 24 h as described above. At the end of each experiment the cells were washed in PBS, fixed in 3.7% formaldehyde, and permeabilized with 0.1% Triton-X and 0.1% sodium citrate, and the incidence of apoptosis was determined by terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick-end labeling (TUNEL) with the TACS TdT fluorescent in situ apoptosis detection kit (R&D Systems) according to the manufacturer's suggestions. Apoptotic nuclei were evaluated by enumerating the total number of FITC-labeled nuclei (excitation and transmission wavelengths 495 and 525 nm, respectively) per microscopic field (×200). Sections were counterstained with the nucleic acid binding dye 4′,6-diamidino-2-phenylindole (Vector Labs; excitation and transmission wavelengths 310 and 353 nm, respectively) to visualize the entire population of fibroblast nuclei per field. Final values were expressed as the percentage of TUNEL-positive fibroblasts. For each slide we examined a total of 6–10 fields.
Effect of MMP inhibition on cardiac fibroblast proliferation.
Fibroblast cultures were seeded in the 96-well plates at a final volume of 100 μl/well and were treated for 24, 48, or 72 h with diluent (PBS) or 5, 50, or 500 nM RS-130830, a broad-based MMP inhibitor provided by AstraZeneca that inhibits MMP-2, -3, -8, -9, -12, -13, and -16 (IC50 0.5–10 nM). Cell proliferation was measured by BrdU incorporation as described above.
Effect of AdTIMP infection and AdTIMP-conditioned medium on cardiac fibroblast collagen synthesis.
Cardiac fibroblast cultures were treated for 72 h with AdTIMP 1–4, AdNull, or 20 ng/ml human transforming growth factor (TGF)-β (R&D Systems), which was used as a positive control. Total collagen synthesis was measured by the Peterkofsky method, with minor modifications (22, 26). Briefly, cardiac fibroblasts were seeded in 12-well plates and infected with AdTIMP-1, -2, -3, -4 or AdNull for 48 h. After 2 days, 1 ml of serum-free medium plus 1 μl of 0.5 M ascorbic acid (Sigma) was added to each well for 10 min at 37°C, followed by labeling of the cells with 5 mCi of l-[2,3,4,5-3H]proline (Amersham Biosciences) for 24 h. After 24 h, the cells and conditioned media were collected and processed exactly as described previously (22, 26). Each sample was split equally into two tubes, A and B, containing 25 μl of 25 mM CaCl2 and 25 μl of 62.5 mM N-ethylmaleimide. Sample A was treated with 25 μl of 1 mg/ml bacterial collagenase-D (Sigma), whereas sample B was treated with 25 μl Krebs-Henseleit buffer. Both samples were then incubated at 37°C for 1.5 h. After incubation, 500 μl of 10% TCA and 10 mM collagen were added to each sample and incubated on ice for an additional 10 min. The samples were then centrifuged at 5,000 rpm for 5 min at 4°C, and the resultant supernatant (600 μl) was added to a scintillant, followed by liquid scintillation counting. To determine the amount of [3H]proline specifically incorporated into collagen, the difference between sample A and sample B was determined. Final results were expressed in terms of counts per minute per sample.
Effect of AdTIMP infection on cardiac fibroblast α-smooth muscle actin expression.
Cardiac fibroblast cultures were infected with AdTIMP-1, -2, -3, -4, or AdNull or treated with 20 ng/ml human TGF-β (R&D Systems), which was used as a positive control. After 72 h, the cells were harvested with trypsin-EDTA (0.05% trypsin, 0.53 mM EDTA) and fixed in 2% paraformaldehyde. The cells were then permeabilized with 0.1% digitonin (Sigma) and incubated with FITC-conjugated mouse monoclonal anti-α-smooth muscle actin (SMA) antibody (Sigma) for 30 min. Cells were washed with PBS and analyzed on a Coulter Epics XL-MCL flow cytometer (Beckman Coulter, Fullerton, CA) with 488-nm excitation and a 530/30-nm band-pass filter for FITC. A FITC-conjugated IgG2a (Sigma) antibody was used to set the acquisition threshold at 1–2% for nonspecific binding to the cardiac fibroblasts. Over 5 × 104 cells were acquired for each sample. We showed previously (37) that the α-SMA antibody used in these experiments labels myofibroblasts in culture.
Each value is expressed as mean ± SE. A one-way ANOVA was used to test for mean differences in BrdU incorporation and collagen synthesis. Where appropriate, post hoc ANOVA testing (Tukey's) was used to assess mean differences between groups at a given time point. A P value <0.05 was considered significant.
Characterization of AdTIMPs in cardiac fibroblasts.
Figure 1A shows that TIMP-1, -2, -3, and -4 protein levels increased with increasing MOI of AdTIMP-1, -2, -3, and -4, with maximal TIMP expression observed at an MOI of 2,500. Figure 1B depicts the time course of TIMP secretion at an MOI of 2,500. As shown, TIMP-1, -2, -3, and -4 levels were undetectable in the medium at 24 h but were easily detectable 72 h after adenoviral infection. On the basis of the foregoing preliminary control experiments, we examined the effects of AdTIMP-1, -2, -3, and -4 (MOI 2,500) at 72 h. In preliminary control experiments we established that the rates of cell proliferation were similar in AdNull-infected cardiac fibroblasts and diluent-treated controls (data not shown).
Effect of AdTIMP infection on cardiac fibroblast proliferation.
Figure 2 shows that infection with AdTIMP-1, -2, -3, or -4 resulted in a significant (P < 0.0001) increase in BrdU incorporation in the cardiac fibroblast cultures compared with AdNull-infected cells. Infection with AdTIMP-1, -2, and -4 resulted in a ∼1.3-fold increase in BrdU incorporation, whereas infection with AdTIMP-3 increased BrdU incorporation ∼1.9-fold compared with AdNull-infected control cells. There was no significant difference in BrdU incorporation among TIMP-1, -2, and -4; however, BrdU incorporation was significantly greater with TIMP-3 (P < 0.001) than with TIMP-1, -2, and -4.
Effect of AdTIMP-conditioned media on cardiac fibroblast proliferation.
To confirm the effects of AdTIMP-1, -2,-3, and -4 infection on fibroblast growth, we examined the effect of AdTIMP-1, -2, -3, and -4 conditioned media on cardiac fibroblast proliferation. The salient finding shown by Fig. 3 is that 24 h of stimulation with TIMP-1-, -2-, -3-, or -4-conditioned medium provoked a significant (P < 0.0001) increase in BrdU incorporation. As shown, TIMP-1, -2, -3, and -4 each increased BrdU incorporation ∼1.2-fold compared with conditioned medium from fibroblasts infected with AdNull virus. There was, however, no significant difference observed in BrdU incorporation among fibroblast cultures that had been stimulated with AdTIMP-1, -2, -3, or -4. The specificity of the effects of TIMP-1, -2, -3, and -4 on BrdU incorporation in cardiac fibroblast was confirmed with a neutralizing anti-TIMP-1, -2, -3, or -4 antibody, which completely blocked the effects of TIMP-1, -2, -3, or -4 on cardiac fibroblast cell proliferation (Fig. 3). There was no significant difference (P > 0.05) between cardiac fibroblasts treated with AdNull-conditioned medium and AdTIMP-1-, -2-, -3-, and -4-conditioned media treated with neutralizing anti-TIMP antibodies.
Effect of AdTIMP infection and AdTIMP-conditioned media on cardiac fibroblast apoptosis.
As shown in Table 1, infection with AdTIMP-1, -2, and -4 had no effect on the incidence of apoptosis in cardiac fibroblasts relative to AdNull-infected fibroblasts, whereas infection with AdTIMP-3 resulted in ∼13-fold increase in apoptosis compared with control values. To determine whether the increased apoptosis in AdTIMP-3-infected cells was secondary to an adenovirus-mediated effect, we also examined the effect of AdTIMP-3-conditioned medium on fibroblast apoptosis. These studies showed that there was a significant increase (P < 0.001) in the incidence of apoptosis with AdTIMP-3-conditioned medium (1.25 ± 0.14%) compared with AdNull-conditioned medium (0.34 ± 0.67%). The slightly lower rate of fibroblast apoptosis observed with conditioned medium compared with the rate observed with AdTIMP-3 infection may have been related to the shorter duration of exposure (24 h vs. 72 h) or may be related to the fact that TIMP-3 adheres to the extracellular matrix and was thus less available for cell stimulation (4). In preliminary control experiments we established that the rates of fibroblast apoptosis were similar in AdNull-infected cardiac fibroblasts and diluent-treated controls (data not shown).
Effect of MMP inhibition on cardiac fibroblast proliferation.
Given that MMP inhibition may stimulate cell proliferation by stabilizing cell surface receptors, as well as by stabilizing the expression of growth factors on the cell membrane, we sought to determine whether the effects of AdTIMP-1, -2, -3, and -4 on fibroblast proliferation were mediated indirectly through inhibition of MMPs secreted by cardiac fibroblasts. Accordingly, cardiac fibroblast cultures were treated for 72 h with 5, 50, and 500 nM RS-130830, a broad-based MMP inhibitor. Figure 4 shows that MMP inhibition had no significant (P > 0.05) effect on fibroblast proliferation at 72 h. Moreover, treatment with RS-130830 had no significant effect on cell proliferation at 24 or 48 h (data not shown). Thus these results suggest that TIMP-1, -2, -3, and -4 stimulate cardiac fibroblast proliferation though a MMP-independent mechanism.
Effect of AdTIMP infection and AdTIMP-conditioned media on cardiac fibroblast collagen synthesis.
To determine whether TIMP-1, -2, -3, and -4 have an effect on collagen synthesis, [3H]proline incorporation was examined in AdTIMP-1-, -2-, -3-, and -4-infected cultures. In parallel experiments, fibroblast cultures were also treated with AdTIMP-1, -2, -3, and -4 conditioned media. Figure 5A shows that AdTIMP-1, -3, and -4 had no significant effect on collagen synthesis. In contrast, AdTIMP-2 infection significantly increased cardiac fibroblast collagen synthesis compared with AdNull-infected cells (P < 0.0001). Figure 5B shows that conditioned media (24 h) from AdTIMP-1-, -3-, and -4-infected cultures had no significant effect on collagen synthesis, whereas AdTIMP-2-conditioned medium significantly (P < 0.001) increased cardiac fibroblast collagen synthesis. The effects of TIMP-2 on cardiac fibroblast collagen synthesis were not significantly different (P > 0.05) from those observed with TGF-β stimulation. The specificity of the effects of TIMP-2 on collagen synthesis in cardiac fibroblasts was confirmed with a neutralizing anti-TIMP-2 antibody, which reduced collagen synthesis ∼1.25-fold (P < 0.0025) compared with TIMP-2-conditioned medium (data not shown). Importantly, the extent of collagen synthesis in the anti-TIMP-2 antibody-TIMP-2-conditioned medium was not significantly (P > 0.05) different from control values.
Effect of AdTIMP infection on cardiac fibroblast α-SMA expression.
To determine whether TIMP-1, -2, -3, and -4 have an effect on the phenotypic differentiation of cardiac fibroblasts into myofibroblasts, we performed fluorescence-activated cell sorting analysis on AdNull- and AdTIMP-1-, -2-, -3-, and -4-infected cardiac fibroblasts. As shown in Fig. 6, infection with AdTIMP-1, -2, -3, or -4 resulted in a significant (P < 0.015) 1.45- to 1.54-fold increase in the frequency of α-SMA-staining fibroblasts relative to AdNull-treated cells. By way of comparison, treatment for 72 h with TGF-β resulted in an approximately twofold increase in the frequency of α-SMA-staining fibroblasts. To determine whether there was heterogeneity within the subpopulation of AdTIMP-1-, -2-, -3-, and -4-infected fibroblasts for α-SMA+ fibroblasts, we also quantified the amount of α-SMA staining in each subpopulation. As shown in Table 2, there was a significant 1.2- to 1.3-fold increase in the extent of α-SMA staining in the AdTIMP-2- and AdTIMP-4-infected cells relative to AdNull-infected cells, whereas the extent of α-SMA staining was not significantly different in the AdTIMP-1- and AdTIMP-3-infected cultures.
The results of the present study, in which we examined the effects of TIMP-1, -2, -3, and -4 on cardiac fibroblasts, show that TIMPs exert overlapping, as well as divergent, effects on cardiac fibroblasts (summarized in Table 3). The following lines of evidence support this statement. First, when cardiac fibroblasts were infected with AdTIMP-1, -2, -3, and -4 constructs, we observed a significant increase in BrdU incorporation, suggesting that TIMP-1, -2, -3, and -4 provoked increased fibroblast proliferation, consistent with previous reports that TIMPs are capable of stimulating cell growth (4). This finding was unlikely to be secondary to nonspecific adenovirus-mediated effects, insofar as AdTIMP-1-, -2-, -3-, and -4-conditioned media also provoked a significant effect on fibroblast proliferation that was similar in magnitude to the effects observed with AdTIMP-infected fibroblasts (Fig. 3) and was completely abolished by treating the AdTIMP-conditioned medium with anti-TIMP-1, -2, -3, and -4 antibodies. Although the results of the studies that used AdTIMP constructs suggested that AdTIMP-3 provoked a greater increase in fibroblast proliferation than AdTIMP-1, -2, and -4, we did not observe a similar increase in the studies using conditioned medium. Given that apoptotic fibroblasts incorporate BrdU (34), it is possible that the increased in BrdU uptake in TIMP-3-infected cells, relative to TIMP-1-, -2-, and -4 infected cells, may have been due, at least in part, to increased apoptosis in the TIMP-3-infected cells (Table 1). The results of the present study further suggest that the effects of TIMP-1, -2, -3, and -4 were receptor mediated, rather than MMP-mediated, insofar as a broad-based MMP inhibitor had no effect on fibroblast growth. Nonetheless, one important limitation of the present study is that we have not identified which fibroblast receptors are responsible for the effects of TIMPs-1, -2, -3, and -4. Indeed, given the breadth of the receptor-mediated TIMP effects (7, 15, 28), it is likely that these molecules engage a variety of different fibroblast receptors.
A second interesting and unanticipated finding was that there was heterogeneity with respect to the effects of TIMPs on fibroblast activation. That is, whereas TIMP-1, -2, -3, and -4 each led to an increase in α-SMA labeling (Fig. 6B), TIMP-2 provoked the greatest increase in α-SMA expression within the subpopulation of α-SMA+ fibroblasts (Table 2). Moreover, both AdTIMP-2-infected fibroblasts and AdTIMP-2-conditioned medium provoked a significant four- to fivefold increase in collagen synthesis (Fig. 5), whereas AdTIMP-1-, -3-, and -4-infected fibroblasts (Fig. 5A) and AdTIMP-1-, -3-, and -4-conditioned medium (Fig. 5B) had no significant effect on fibroblast collagen synthesis. Finally, as noted above, TIMP-3 provoked an increase in fibroblast apoptosis (Table 1), consistent with previous reports in smooth muscle cells (9), whereas TIMP-1, -2, and -4 had no effect on fibroblast apoptosis.
Role of TIMPs in the heart.
In experimental models of chronic injury and/or inflammation in the heart, liver, lung, and kidney there is an initial increase in MMP expression that is invariably superseded by increased expression of fibrogenic cytokines (e.g., TGF-β), as well as increased expression of TIMPs (10, 19, 30). It has been suggested that time-dependent changes in the balance between MMP activity and TIMP levels in the heart play a critical role in the process of left ventricular (LV) remodeling (13, 21, 31, 32). That is, whereas an increase in the ratio of MMP activity to TIMP levels favors matrix degradation and LV dilation, a decrease in this ratio is associated with stabilization of LV dilation and progressive myocardial fibrosis (31). Consistent with this point of view, TIMP-1-knockout mice exhibit decreased myocardial fibrillar collagen and increased LV dilation and hypertrophy after acute myocardial infarction (12), whereas TIMP-1 administration at the time of infarction prevents LV rupture (17), suggesting that TIMP-1 plays an important role in the process of LV remodeling after tissue injury. Subsequent studies have demonstrated that TIMP-1 and -2 mRNA and protein and TIMP-3 mRNA levels are increased after acute myocardial infarction (27, 29, 33), in keeping with a potential role for these molecules in governing tissue repair after cardiac injury. Although the canonical role that has been assigned to TIMPs is that of inhibiting the active forms of MMPs, a growing body of evidence suggests that TIMPs exert effects that cannot be ascribed to their ability to attenuate active MMPs. Indeed, TIMPs have been shown to both stimulate and inhibit cell growth, prevent angiogenesis, as well as regulate cell survival and cell apoptosis (2, 11, 18, 24, 38, 39). Although TIMP-1 and -3 and TIMP-2 and -4 share sequence homologies, the reasons for the significant functional differences among TIMP-1, -2, -3, and -4 are not known, underscoring the need for mapping the regions of TIMPs that are responsible for these activities to better understand the novel structure-function relationships of these molecules.
Relevant to the above discussion, the results of the present study suggest that TIMPs may contribute importantly to tissue repair through mechanisms that are independent of their known ability to inactivate MMPs. For example, in addition to inactivating MMPs, TIMP -1, -2, -3, and -4 might stimulate fibroblast proliferation as well as phenotypic differentiation into myofibroblasts at the site of tissue injury. This in turn might lead to improved scar formation, by virtue of the ability of the myofibroblasts to align and contract the extracellular matrix. Furthermore, increased expression of TIMP-2 could contribute importantly to wound healing by stimulating increased fibroblast collagen synthesis within the microenvironment of tissue injury. Finally, increased expression and localization of TIMP-3 within at the site of tissue injury might facilitate programmed cell death of injured and/or damaged cardiac myocytes and/or contribute to the death of activated myofibroblasts, thus acting as an “off switch” that prevents and/or delimits excessive myocardial fibrosis. It bears emphasis, however, that these suggestions remain speculative and that additional in vivo studies will be required to test these possibilities. The present report differs from a previous study in which TIMP-4 purified from human hearts had no effect on cardiac fibroblast proliferation whereas TIMP-4 inhibited cell growth and provoked apoptosis of transformed cardiac fibroblasts (35). Although the reasons for this discrepancy are not known, they may relate to differences in culture conditions, species differences, differences in the versions of TIMP-4 that were used (i.e., purified vs. genetically expressed), or differences in the levels of TIMP expression.
Limitations of present study.
Although the precise local concentration of TIMPs at the site of tissue injury is not known, we cannot exclude the possibility that the concentrations of TIMPs used in the present study are supraphysiological. However, it bears emphasis that the results of the present study are consistent with previous studies in the literature that have used lower MOIs for AdTIMP infection (1, 5). Another limitation of the present study is that we did not determine the mechanism(s) for TIMP-3-mediated cardiac fibroblast apoptosis. However, we showed previously (2, 8, 9) that TIMP-3 sensitizes cells to death receptor-mediated apoptosis, that TIMP-3-mediated apoptosis is sensitive to inhibition with a broad-spectrum caspase inhibitor (Z-VAD) and/or bcl-2, and that TIMP-3-mediated apoptosis can be inhibited by crmA and/or a dominant-negative FADD construct.
In conclusion, to our knowledge these studies comprise the first systematic study of the effects of TIMP-1, -2, -3, and -4 on cardiac fibroblasts (summarized in Table 3) and constitute the initial demonstration that TIMPs provoke a phenotypic switch in cardiac fibroblasts to a more activated myofibroblast phenotype that, at least in the case of TIMP-2, is also capable of increased collagen synthesis. As noted above, these observations suggest that TIMPs may play an important role in tissue injury in the heart that extends beyond the traditional role that has been ascribed to these molecules as natural inhibitors of MMPs. Apart from the novelty of these findings, these studies raise the intriguing possibility that the biology of TIMPs might be exploited therapeutically. For example, TIMPs (e.g., TIMP-1, -2, -3, -4) that have little profibrotic activity in cardiac fibroblasts might be used to alter an unfavorable MMP-TIMP balance in hearts that are undergoing progressive cardiac remodeling, perhaps with gene-based approaches. Alternatively, TIMPs that have increased profibrotic activity (e.g., TIMP-2) might be introduced mechanically at the site of tissue injury to facilitate wound healing. Ongoing studies are being conducted to address these interesting possibilities.
This research was supported by research funds from the Department of Veterans Affairs and the National Heart, Lung, and Blood Institute (P50-HL-O6H and RO1-HL-58081-01, RO1 HL-61543-01, and HL-42250-10/10). J. D. Lovelock was a research scholar supported by the Stanley J. Sarnoff Endowment for Cardiovascular Science, Inc.
The authors gratefully acknowledge the technical assistance of Dorellyn Lee Jackson and the secretarial support of Mary Helen Soliz.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 by the American Physiological Society