Atrial but Not Ventricular Fibrosis in Mice Expressing a Mutant Transforming Growth Factor-β1 Transgene in the Heart
Abstract—Increased transforming growth factor (TGF)–β1 activity has been observed during pathologic cardiac remodeling in a variety of animal models. In an effort to establish a causal role of TGF-β1 in this process, transgenic mice with elevated levels of active myocardial TGF-β1 were generated. The cardiac-restricted α–myosin heavy chain promoter was used to target expression of a mutant TGF-β1 cDNA harboring a cysteine-to-serine substitution at amino acid residue 33. This alteration blocks covalent tethering of the TGF-β1 latent complex to the extracellular matrix, thereby rendering a large proportion (>60%) of the transgene-encoded TGF-β1 constitutively active. Although similar levels of active TGF-β1 were present in the transgenic atria and ventricles, overt fibrosis was observed only in the atria. Surprisingly, increased active TGF-β1 levels inhibited ventricular fibroblast DNA synthesis in uninjured hearts and delayed wound healing after myocardial injury. These data suggest that increased TGF-β1 activity by itself is insufficient to promote ventricular fibrosis in the adult mouse ventricle.
Pathologic cardiac remodeling refers to the process by which the normal cardiac architecture is altered in response to myocardial infarction, ischemia, and/or cardiomyocyte loss.1 Remodeling is frequently accompanied by hypertrophic growth of cardiac myocytes and hyperplastic growth of cardiac fibroblasts. There is also an increased deposition of extracellular matrix constituents (ie, fibrosis). It is generally believed that increased extracellular matrix content contributes to diastolic stiffness and that this process ultimately promotes ventricular dysfunction. Myocardial fibrosis is also a major risk factor for arrhythmia susceptibility.
Regulation of cardiac fibroblast proliferation and collagen synthesis is thought to be multifactorial in nature.1 2 The renin-angiotensin-aldosterone system, transforming growth factor (TGF)–β family members, acidic and basic fibroblast growth factor, endothelins, and tumor necrosis factor-α have all been implicated in these processes. Dysregulation of TGF-β1 is thought to promote fibrosis in a number of organs, and in humans the lung, liver, and kidney appear to be particularly sensitive.3 Circumstantial evidence supports a potential role for TGF-β in cardiac fibrosis. For example, increased TGF-β1 mRNA has been observed in hypertrophic hearts,4 and a marked increase in transcript levels accompanied the onset of heart failure in aged spontaneously hypertensive rats.5 6 In vitro studies indicated that TGF-β1 stimulates extracellular matrix deposition from cultured cardiac fibroblasts.7 8 9 These observations collectively suggest a causal role for TGF-β in cardiac fibrosis.
TGF-β1 and its closely related homologues (TGF-β2 and TGF-β3) are members of a large superfamily of cytokines that regulate the proliferation, differentiation, migration, and viability of a large variety of cell types.10 TGF-β1 is a homodimer that is secreted as a large inactive (or latent) multiprotein complex that is subsequently tethered to the extracellular matrix. Recent studies indicate that activation of TGF-β1 requires proteolytic release of the multiprotein complex from the extracellular matrix,11 followed by hydrostatic binding with activating proteins. Binding of TGF-β1 to activating proteins induces a conformational change, thereby permitting interaction between the mature TGF-β1 homodimer and its cognate receptor complex.12 TGF-β1 signals through a heteromeric complex comprising 2 serine threonine kinase receptors (TβR1 and TβR2). TGF-β1 binding to TβR2 promotes phosphorylation of the TβR1 receptor, which in turn propagates the signal to downstream intracellular targets.13
To ascertain the role of TGF-β1 in induction of cardiac fibrosis, we have generated transgenic mice that express a mutated human TGF-β1 cDNA under the transcriptional regulation of the α–cardiac myosin heavy chain (MHC) promoter. The mutation encompasses a cysteine-to-serine substitution at amino acid residue 33, which blocks covalent tethering of the TGF-β1 latent complex to the extracellular matrix. Transgene expression resulted in a marked increase in active TGF-β1 levels in adult hearts. A fibrotic response was observed in the atria of the transgenic mice, but surprisingly not in the ventricles. Further characterization of normal and injured mice revealed that transgene expression inhibited ventricular fibroblast DNA synthesis and delayed myocardial wound healing. The data suggest that in vivo, TGF-β1 expression may be insufficient to promote ventricular fibrosis. The results are discussed within the context of the known effects of TGF-β1 on cultured cardiac fibroblasts.
Materials and Methods
Generation of the MHC-TGFcys33ser Transgene Mice
The MHC-TGFcys33ser transgene used the mouse α–cardiac MHC promoter14 and sequences encoding the human TGF-β1 cDNA (see Figure 1A⇓). The simian virus 40 (SV40) early region transcription terminator/polyadenylation site (nucleotide residues 2586 to 245215 ) was inserted downstream from the TGF cDNA. Transgene insert was purified and microinjected into inbred C3HeB/FeJ (The Jackson Laboratories, Bar Harbor, ME) zygotes, which were then implanted into pseudopregnant mice as described.16 The resulting pups were screened using diagnostic polymerase chain reaction (PCR) amplification.17 For all surgeries, mice were anesthetized with 2.5% tribromoethanol in amylene hydrate (Avertin; 0.015 mL/g body weight IP, Fluka Biochemicals). All manipulations were performed according to NIH and Institutional Animal Care and Use Guidelines.
Northern Blot and Reverse Transcription–PCR Analyses
RNA was prepared by extraction with guanidinium thiocyanate and analyzed (10 μg) by Northern blotting as described.17 18 Probes were radiolabeled by nick translation and hybridized using standard protocols.18 For quantitative reverse transcription (TaqMan)–PCR analysis, 1 μg of total RNA was used to prime the cDNA synthesis reaction according to the manufacturer’s instructions (Promega reverse transcription system, model A3500). TaqMan PCR analysis was then performed using the Perkin Elmer PCR core reagent kit (N808-0228) and Applied Biosystems SDS as detection system. The primer and probe combinations were obtained from published sequences for mouse α1 collagen type I (GenBank accession No. U08020), mouse TGF-β1 (GenBank accession No. M13177), mouse lysyloxidase (GenBank accession No. M65142), mouse plasminogen activator inhibitor 1 (PAI-1; GenBank accession No. M33960), and mouse hypoxanthine phosphoribosyl transferase (GenBank accession No. J00423). Reaction products were quantified according to the Delta-Delta threshold cycle method. The values presented thus reflect the fold induction of marker gene expression in the transgenic hearts as compared with the nontransgenic controls.
Cryosections (10 μm) were generated using standard histological techniques.21 Paraffin sections were prepared from Bouin’s fixed samples using standard procedures.21 Hematoxylin and eosin (H&E), Sirius red, and Masson’s trichrome staining were all performed according to manufacturer specifications (Sigma Diagnostics). For minimal fiber analyses, images were captured, digitized, and analyzed with NIH Image 1.61 software as described.22 Images from Masson’s trichrome–stained sections were captured and the collagen content determined by quantifying the blue pixel content using Adobe Photoshop 5.02 software. For terminal deoxynucleotidyltransferase–mediated dUTP nick-end labeling (TUNEL) analysis of DNA fragmentation, cryosections were processed using the Boehringer Mannheim in situ cell death detection kit.
Myocardial Damage Model
Cautery injury was performed in an open chest protocol as described previously.22 The mortality rate for the procedure was <5%.
Fibroblast DNA Synthesis Assays
The statistical significance was assessed by the Student unpaired t test. Probability values <0.05 were considered as statistically significant. All results are presented as mean±SEM.
An expanded Materials and Methods section is available online at http://www.circresaha.org.
Generation of Mice With Elevated Myocardial Levels of Active TGF-β1
Post-transcriptional regulation of TGF-β1 is complex. The structure of preproTGF-β1 and the extracellular matrix–associated large latent form of the molecule are depicted in Figure 1A⇑ and 1B⇑, respectively. Mature TGF-β1 (comprising amino acid residues 279 to 390 of preproTGF-β1) is normally secreted as a biologically inactive disulfide-linked homodimer. Inactivity results from noncovalent binding of the mature TGF-β1 homodimer to a disulfide-linked homodimer of the latency-associated peptide (LAP), which is generated by proteolytic processing of amino acid residues 31 to 278 of preproTGF-β1. This structure is known as the small latent complex (Figure 1B⇑). The small latent complex is tethered to the extracellular matrix via covalent bond formation between the LAP dimer and a large, alternatively spliced glycoprotein called latent TGF binding protein 1 (LTBP-1). The resulting structure is known as the large latent complex (Figure 1B⇑). Release of biologically active TGF is thought to entail proteolytic cleavage of the small latent complex from extracellular matrix, followed by dissociation of mature TGF-β1 from the LAP.
Given this complex post-transcriptional regulation, several strategies using site-directed mutagenesis of the LAP have been developed to facilitate secretion of constitutively active TGF-β1. Our studies used a TGF-β1 cDNA encoding a cysteine-to-serine substitution at TGF-β1 amino acid residue 33, which prevented covalent binding between the small latent complex and LTBP-1, thereby blocking formation of large latent complex.11 COS cells transfected with this mutant exhibited greater TGF activity as compared with cells transfected with wild-type constructs.24
A transgene comprising the MHC promoter and a mutant human TGF-β1 cDNA encoding the residue 33 mutation was generated (designated MHC-TGFcys33ser). PCR analyses with oligonucleotide primers specific for the MHC-TGFcys33ser sequences revealed that 6 of the 36 mice derived from microinjected embryos were transgenic. Five of these animals successfully transmitted their transgenes to the F1 generation. Northern blot analysis was used to stratify the level of recombinant TGF-β1 expression in these 5 transgenic lineages (Figure 1D⇑). High levels of transgene-derived TGF-β1 mRNA were detected in MHC-TGFcys33ser lines 3 and 4, intermediate levels were observed in lines 5 and 6, and no transgene expression was evident in line 2 (transgene-encoded TGF-β1 transcripts reproducibly run as a doublet in the glyoxal gels used). MHC-TGFcys33ser line 3 was selected for subsequent experiments. Northern blot analyses revealed similar levels of transgene-encoded TGF-β1 transcripts in the atria and ventricles of adult mice (Figure 1D⇑). Similar levels of transgene mRNA were detected in young (6 weeks) and old (12 months) MHC-TGFcys33ser mice (not shown).
Transgene Expression and Basic Cardiac Attributes in MHC-TGFcys33ser Mice
To confirm that transgene expression in the MHC-TGFcys33ser mice resulted in increased TGF-β activity, heart homogenate was prepared from 11-week-old male and female MHC-TGFcys33ser line 3 transgenic mice and from their nontransgenic siblings. The homogenate was tested directly in a mink lung cell growth bioassay to quantify the levels of active TGF-β in the samples. There was ≈30-fold more active TGF-β in the transgenic hearts as compared with the controls (Figure 1E⇑). To determine the total TGF-β content, the heart homogenates were acidified (which disrupts LAP/TGF-β binding, thereby releasing all of the mature TGF-β from the latent complex). Total TGF activity was only ≈1.3-fold greater in the transgenic hearts as compared with the controls (Figure 1E⇑). Latent TGF-β content was calculated as the difference between total TGF-β activity (ie, the value from the acid-treated samples) and active TGF-β (ie, the value from the nontreated samples). The preponderance (>60%) of the TGF-β in the transgenic hearts was in the active form, as compared with only 3% for the nontransgenic controls (Figure 1E⇑). Although transgene expression resulted in constitutively high levels of active TGF-β1 in adult transgenic hearts, this activity was less than the total reserve activity (ie, active plus latent TGF-β1) present in nontransgenic hearts. Thus, if appropriately stimulated, a nontransgenic heart could produce more TGF-β activity than that which was constitutively active in the MHC-TGFcys33ser transgenic hearts. Finally, TGF-β bioassay revealed similar levels of transgene activity in the transgenic atria and ventricles (data not shown).
Basic cardiac attributes were monitored to determine whether expression of the MHC-TGFcys33ser transgene had an adverse effect on the transgenic hearts. A slight but statistically significant increase in heart weight was apparent in the transgenic mice as compared with sex- and aged-matched siblings (Table 1⇓; 11-week-old male mice were analyzed). This increase was also apparent when the heart weights were normalized to tibia length. The mean minimal fiber diameter of the transgenic ventricular cardiomyocytes was also increased in the transgenic hearts, suggesting that the elevated heart weight resulted from TGF-β1–induced myocardial hypertrophy (as opposed to an increase in ventricular cardiomyocyte number). Interestingly, the minimal fiber diameter of the transgenic atrial cardiomyocytes was less than that for the control animals, suggesting that transgene expression had a differential effect on atrial versus ventricular cardiomyocytes.
Elevated Myocardial TGF-β1 Activity Promotes Atrial but Not Ventricular Fibrosis
Histological samples from line 3 hearts were examined to determine whether sustained elevation of active TGF-β1 levels promoted cardiac fibrosis. Surprisingly, H&E analyses failed to detect any overt differences between control and transgenic hearts (Figure 2A⇓ and 2D⇓, respectively). Additional analyses with Masson’s trichrome (which stains collagen blue) also failed to detect differences between the control and transgenic ventricles, although collagen deposition was apparent in the transgenic atria (Figure 2E⇓). The atrial histopathology was further characterized using Sirius red staining with a fast green counterstain: intense red staining indicative of cardiac fibrosis was readily apparent in the transgenic atria, but not in age-matched control atria (Figure 2F⇓ and 2C⇓, respectively).
Atrial fibrosis was age dependent in the MHC-TGFcys33ser mice. No atrial histopathology was apparent in Sirius red–stained sections prepared from 4-week-old transgenic hearts (Figure 2G⇑). Fibrosis was abundant at 20 weeks of age (Figure 2H⇑) and was increased in severity at 8.5, 18.5, and 23 months of age (Figure 2I⇑ through 2K, respectively). In contrast, no evidence of fibrosis was apparent in the ventricles of the same sections (a portion of the ventricles are visible in Figure 2G⇑ through 2K), nor in the atria of age-matched nontransgenic siblings (not shown). Because atrial fibrosis was observed in 3 independent MHC-TGFcys33ser transgenic lineages, this phenotype did not result from a transgene insertional mutation (MHC-TGFcys33ser lines 3, 4, and 6 were examined).
To further explore the differential pathophysiology seen in MHC-TGFcys33ser line 3 mice, RNA prepared from the atrial and ventricular myocardium was examined for the expression of several genes induced during fibrosis. TaqMan reverse transcription–PCR analyses indicated that expression of α1 collagen type 1, lysyloxidase, and PAI-1 were markedly elevated in atria of MHC-TGFcys33ser mice as compared with nontransgenic atria (Figure 3⇓), in good agreement with the histochemical data presented above. In addition, expression of the endogenous TGF-β1 gene was also elevated (the probe used for TGF-β1 quantification did not react with the human gene product encoded by the transgene). In contrast, essentially no increase in fibrosis-related gene expression was seen in RNA prepared from the ventricles of the MHC-TGFcys33ser mice as compared with control samples, in agreement with the histochemical data.
Elevated Myocardial TGF-β1 Is Associated With Delayed Wound Healing in the Ventricle
The presence of atrial, but not ventricular, fibrosis in the MHC-TGFcys33ser mice was somewhat surprising, particularly in light of the similar levels of transgene-encoded TGF-β1 mRNA (Figure 1D⇑) and activity (see above). Additional experiments were initiated to determine whether an exaggerated fibrotic response to myocardial injury would occur in the ventricles of the transgenic mice. Male mice 11 weeks old were analyzed. The left ventricular free wall of transgenic and control mice was injured by cauterization. Previous studies have shown that this procedure produces a necrotic zone at the site of cauterization, as well as an ischemic zone distal to and apically located from the cauterization site.22 A hypertrophic response in the surviving cardiomyocytes is also observed with this injury model. Assessment of minimal fiber diameter 1 week after injury revealed that cardiomyocytes proximal to the injury site had hypertrophied to a similar size in the nontransgenic and MHC-TGFcys33ser mice (Table 2⇓). Although cardiomyocytes distal to the injury site exhibited a mild hypertrophic response, the underlying difference in control versus transgenic minimal fiber diameter seen previously in uninjured hearts was still apparent (Table 2⇓).
Gross analysis of nontransgenic hearts at 1 week after injury revealed the presence of a well-formed fibrous cap over the cautery-induced necrotic zone. H&E analysis of sections from these hearts confirmed that healing of the necrotic zone was well underway (Figure 4A⇓). The epicardial surface was essentially intact, and the necrotic zone contained numerous fibroblasts and leukocytes. Masson’s trichrome staining of an adjacent section revealed significant collagen deposition throughout the damaged area (Figure 4B⇓). Gross examination of hearts from nontransgenic mice at 2 weeks after injury revealed that the fibrous scar had begun to contract, an indication that the healing process was nearly complete. Examination of Masson’s trichrome–stained sections confirmed that the necrotic zone was almost completely healed; widespread and extensive collagen deposition was apparent (Figure 4C⇓).
In contrast, the healing process appeared to be inhibited in the MHC-TGFcys33ser mice. Gross examination of the transgenic hearts at 1 week after injury revealed the presence of ulcerated lesions, indicating that elevated TGF-β1 activity (either directly or indirectly) delayed the healing process. This was confirmed by microscopic analysis of tissue sections; H&E analysis at 1 week after injury revealed that the epicardial surface remained largely disrupted (Figure 4D⇑). Large expanses of the necrotic zone were devoid of fibroblasts and leukocyte infiltration. Analysis of adjacent sections stained with Masson’s trichrome indicated that collagen deposition was also markedly reduced in the transgenic mice (Figure 4E⇑). Gross examinations of transgenic hearts at 2 weeks after injury revealed that although a fibrous capsule was now visible over much of the necrotic zone, scar contraction had not yet occurred. Masson’s trichrome histochemistry indicated that, although significant collagen deposition had occurred in the transgenic hearts by 2 weeks after injury, healing at the epicardial surface was still incomplete. The extent of collagen deposition in the control and transgenic hearts was quantified by image analysis of Masson’s trichrome–stained sections. In agreement with the histological observations, collagen deposition was significantly delayed in the injured transgenic hearts as compared with injured nontransgenic siblings (Table 3⇓).
Elevated TGF-β1 Activity Promotes Fibroblast DNA Synthesis and Apoptosis
The above results suggested that fibroblast migration, growth, and/or survival were inhibited in the MHC-TGFcys33ser transgenic hearts. To monitor fibroblast nuclear density and DNA synthesis, the MHC-TGFcys33ser mice were crossed with MHC-nLAC mice (a transgenic reporter model that permits discrimination between cardiomyocyte and fibroblast nuclei). The MHC-TGFcys33ser/MHC-nLAC (experimental) and MHC-nLAC (control) mice produced by this cross were identified by PCR analysis and sequestered. At 11 weeks of age, the mice received a single injection of [3H]thymidine (both male and female mice were analyzed). The mice were euthanized 4 hours later, and the hearts were harvested and cryosectioned. The resulting sections were processed for X-GAL (5-bromo-4-chloro-3-indolyl β-d-galactosidase) staining (cardiomyocyte nuclei appear dark blue under bright-field illumination), Hoechst staining (cardiac fibroblast nuclei appear light blue under fluorescent illumination), and autoradiography (cells synthesizing DNA have nuclear silver grains). Although similar fibroblast nuclear densities were observed in the MHC-TGFcys33ser and control hearts, the fibroblast [3H]thymidine incorporation indices were ≈3-fold lower in the transgenic animals (Table 4⇓).
This reduction in fibroblast DNA synthesis (and by inference, proliferation) could explain the delayed healing response observed in the MHC-TGFcys33ser hearts. A second series of experiments was performed in cautery-injured hearts to directly test this hypothesis. Once again, MHC-TGFcys33ser/MHC-nLAC experimental and MHC-nLAC control mice were used. At 11 weeks of age, the mice were subjected to cautery injury (both male and female mice were analyzed). After 1 week, the animals received an injection of [3H]thymidine, and the hearts were processed as described above. Similar fibroblast nuclear densities were observed in the injured region bordering intact myocardium in nontransgenic (Figure 5A⇓) and MHC-TGFcys33ser (Figure 5B⇓) hearts. A marked decrease in nuclear density was seen at the subepicardial surface of the damaged region in the transgenic hearts as compared with controls (Figure 5D⇓ and 5C⇓, respectively), consistent with the histochemical analyses presented above. In contrast to the results observed in noninjured hearts, the fibroblast DNA synthesis labeling indices were markedly elevated in the injured region bordering intact myocardium of the MHC-TGFcys33ser hearts as compared with the controls. Elevated fibroblast DNA synthesis labeling indices were also observed at the subepicardial surface of injured MHC-TGFcys33ser and control hearts (see Table 4⇑ for quantification of the fibroblast nuclear densities and labeling indices).
The inconsistency between fibroblast nuclear density and labeling indices suggested that fibroblast survival might be impaired in the injured MHC-TGFcys33ser transgenic hearts. TUNEL staining was used to further explore this possibility. The incidence of TUNEL positivity in the damaged regions of control mice was quite low (Figure 5E⇑), suggesting that fibroblast apoptosis was infrequent in these animals. In contrast, TUNEL positivity was elevated in the injured MHC-TGFcys33ser hearts, frequently with multiple positive cells present in a single microscopic field (Figure 5F⇑). These observations suggest that the impaired healing observed in the MHC-TGFcys33ser hearts may be due, at least in part, to decreased viability of fibroblasts.
The studies presented here demonstrate that targeted expression of a human TGF-β1 cDNA harboring a cysteine-to-serine substitution at amino acid residue 33 of the LAP markedly increased active TGF-β1 levels in adult transgenic hearts. Targeted TGF-β1 expression caused a slight increase in ventricular mass, which appeared to result from a hypertrophic response in the cardiomyocytes. No overt fibrosis was observed in the ventricular myocardium, although age-dependent fibrosis was observed in the atria. A concomitant induction of several fibrosis-related gene products was also observed in the transgenic atria. Transgene expression inhibited ventricular fibroblast DNA synthesis in uninjured hearts and also delayed wound healing after cauterization. The delay in wound healing was due, at least in part, to decreased fibroblast viability in the injured zone of the heart.
As indicated above, cardiac fibrosis is associated with an increase in extracellular matrix content and in particular with an increase in collagen deposition. Several studies have demonstrated that TGF-β1 promotes collagen synthesis and/or deposition in cultured cardiac fibroblasts in vitro.7 8 9 Moreover, concomitant increases in TGF-β1 gene expression have been observed with increased collagen gene expression and deposition in vivo in a variety of animal models.1 2 5 Aside from increased expression of extracellular matrix constituencies, increased TGF-β1 expression is one of the few molecular markers that discriminates between compensated and decompensated cardiac hypertrophy.1 These correlative observations have led to the suggestion that TGF-β1 may play a causative role in cardiac fibrosis.
The absence of overt ventricular fibrosis in the transgenic mice, despite a 30-fold increase in the level of active TGF-β, was thus unexpected. It is important to note that the inhibition of fibroblast DNA synthesis observed in uninjured transgenic hearts is consistent with the previously observed growth-inhibitory effects of TGF-β1 on cultured cardiac fibroblasts in vitro.8 9 Because the dose of TGF-β1 required to elicit fibroblast growth inhibition in vitro was in vast excess of that needed to stimulate collagen synthesis,7 8 it is unlikely that the absence of ventricular fibrosis observed in vivo with the MHC-TGFcys33ser transgenic hearts results from insufficient levels of active TGF-β1. The difference in collagen deposition observed in the in vivo and in vitro models might result from combinatorial effects of growth factors. For example, although fibroblast growth factor by itself can stimulate collagen deposition in cultured cardiac fibroblasts, concomitant treatment with TGF-β1 blocked the induction.25 The complex milieu of growth factors and cytokines present in the in vivo extracellular matrix might mask the stimulatory effects of TGF-β1 on collagen deposition observed in vitro. Alternatively, fibroblasts grown on a 2-dimensional substrate in vitro might display altered responsiveness to cytokines as compared with their in vivo response.
The differential fibrosis observed in the atrial versus ventricular myocardium was surprising, as similar levels of transgene mRNA and TGF-β biological activity were present in both chambers. Several factors might contribute to this phenomenon. For example, there may be a difference in the threshold level of TGF-β1 needed to elicit a response in the atrial versus the ventricular myocardium. This could result from differential expression of requisite ancillary factors (eg, receptors, additional cytokines, or activating proteins). It is also possible that transgene-derived TGF-β1 may have a deleterious effect on atrial cardiomyocytes (but not on ventricular cardiomyocytes) and that the observed atrial fibrosis is a secondary response. Regardless of the molecular basis for the differential response to TGF-β1 expression in this model, it is of interest to note that the atrial histopathology seen in the MHC-TGFcys33ser atria is not dissimilar to the fibrosis observed in some patients with atrial fibrillation.26 27 Atrial hypersensitivity to cytokines might contribute to the etiology of atrial fibrillation in some cases. It is unlikely that the use of a heterologous cDNA contributed to the absence of ventricular fibrosis in our model, as targeted expression of human TGF-β1 transgenes have induced fibrotic responses in other cell types.28 29 Finally, although a slight decrease in atrial cardiomyocyte minimal fiber diameter was noted in the MHC-TGFcys33ser transgenic hearts, it is difficult to gauge the significance of this observation in light of the severe underlying histopathology.
It has recently been reported that targeted cardiac expression of a constitutively activated ALK5 (a TGF type 1 receptor) resulted in embryonic lethality.30 Affected fetuses exhibited linear, dilated hypoplastic heart tubes, and cardiac development arrested before looping. In contrast, MHC-TGFcys33ser embryos were viable, with no apparent cardiac developmental anomalies. The phenotypic differences between these models likely reflects the degree to which the TGFcys33ser ligand versus the constitutive ALK5 receptor are respectively able to activate the TGF-β signal-transduction pathway. In light of this, it is important to reiterate the data indicating that the product of the MHC-TGFcys33ser transgene is biologically active. The mink lung cell growth inhibition bioassay indicated that the level of active TGF-β was 30-fold greater in the transgenic hearts as compared with the control animals. Moreover, the appearance of atrial fibrosis, the general inhibition of ventricular fibroblast DNA synthesis in uninjured hearts, and the impairment of wound healing after cautery injury all indicate that the transgene product was biologically active. It is nonetheless important to acknowledge that any given genetic modification (as exemplified by the targeted expression of active TGF-β1 in this model) has the potential to alter the normal developmental program. Thus, it is formally possible that some or all of the phenotypes attributed to elevated TGF activity in our model might not faithfully mimic the response that would be observed in a genetically naive animal. However, it is also important to note that no overt fibrosis was observed in the myocardium of hearts engrafted with skeletal myoblasts expressing high levels of TGF-β1 encoding the Cys223,225Ser mutation.31 In light of these collective observations, the most direct interpretation of the results obtained here is that expression of active TGF-β1 per se is insufficient to promote ventricular fibrosis in adult transgenic hearts.
The effects of transgene expression on cardiac fibroblast proliferation and apoptosis provide a potential explanation for the delayed wound healing observed in the MHC-TGFcys33ser transgenic mice after cautery injury. The time course of fibroblast accumulation in the injured zones was delayed in the transgenic animals as compared with the nontransgenic controls. Paradoxically, the fibroblast thymidine incorporation indices were elevated. This is in contrast to in vitro experiments in which exogenous TGF-β1 was shown to inhibit cardiac fibroblast proliferation.8 32 However, Agocha et al9 noted that when cultured under hypoxic conditions, TGF-β1 promotes fibroblast proliferation in vitro. Because the cautery injury model induces regional hypoxia, increased fibroblast thymidine incorporation indices are consistent with the in vitro data. Thus, the increased [3H]thymidine incorporation (as well as the delayed collagen deposition; see above) observed in MHC-TGFcys33ser hearts is likely to result from combinatorial effects between the transgene-encoded active TGF-β1 and the complex environment resulting from myocardial injury. These factors undoubtedly also contribute to the apparent increase in fibroblast apoptosis observed in this model.
In summary, the results presented here indicate that a sustained increase in the level of TGF-β1 activity in itself does not promote fibrosis in the ventricular myocardium of transgenic mice. In contrast, fibrosis was observed in the atrial myocardium of the transgenic animals, suggesting that atrial fibroblasts and/or cardiomyocytes exhibit an exaggerated response to this cytokine. Additional studies are required to ascertain what role (if any) TGF-β1 plays in the onset of ventricular fibrosis in humans and, furthermore, to what extent TGF-β1 contributes to pathogenesis in the atrial myocardium.
This work was supported by the National Heart, Lung, and Blood Institute (to L.J.F.) and a Grant-in-Aid from the American Heart Association, Indiana Affiliate (to H.N.). We thank H. Wang and D. Field for excellent technical assistance and Drs M. Soonpaa, L. Pajak, K. Pasumarthi, and S.-C. Tsai for comments on the manuscript.
- Received September 16, 1999.
- Accepted November 29, 1999.
- © 2000 American Heart Association, Inc.
Swynghedauw B. Molecular mechanisms of myocardial remodeling. Physiol Rev. 1999;79:215–262.
Takahashi N, Calderone A, Izzo NJ Jr, Maki TM, Marsh JD, Colucci WS. Hypertrophic stimuli induce transforming growth factor-beta 1 expression in rat ventricular myocytes. J Clin Invest. 1994;94:1470–1476.
Boluyt MO, Bing OH, Lakatta EG. The ageing spontaneously hypertensive rat as a model of the transition from stable compensated hypertrophy to heart failure. Eur Heart J. 1995;16(suppl N):19–30.
Boluyt MO, O’Neill L, Meredith AL, Bing OH, Brooks WW, Conrad CH, Crow MT, Lakatta EG. Alterations in cardiac gene expression during the transition from stable hypertrophy to heart failure: marked upregulation of genes encoding extracellular matrix components. Circ Res. 1994;75:23–32.
Agocha A, Lee HW, Eghbali-Webb M. Hypoxia regulates basal and induced DNA synthesis and collagen type I production in human cardiac fibroblasts: effects of transforming growth factor-beta1, thyroid hormone, angiotensin II and basic fibroblast growth factor. J Mol Cell Cardiol. 1997;29:2233–2244.
Gulick J, Subramaniam A, Neumann J, Robbins J. Isolation and characterization of the mouse cardiac myosin heavy chain genes. J Biol Chem. 1991;266:9180–9185.
Reddy VB, Thimmappaya B, Dhar R, Subramanian KN, Zain BS, Pan J, Ghosh PK, Celma ML, Weissman SM. The genome of simian virus 40. Science. 1978;200:494–502.
Hogan B. Manipulating the Mouse Embryo. Plainview, NY: Cold Spring Harbor Laboratory Press; 1994:497.
Steinhelper ME, Cochrane KL, Field LJ. Hypotension in transgenic mice expressing atrial natriuretic factor fusion genes. Hypertension. 1990;16:301–307.
Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 1989.
Bullock GR, Petrusz P. Techniques in Immunocytochemistry. London, UK: Academic Press; 1982:v.
Soonpaa MH, Field LJ. Assessment of cardiomyocyte DNA synthesis in normal and injured adult mouse hearts. Am J Physiol. 1997;272:H220–H226.
Soonpaa MH, Koh GY, Klug MG, Field LJ. Formation of nascent intercalated disks between grafted fetal cardiomyocytes and host myocardium. Science. 1994;264:98–101.
Brunner AM, Marquardt H, Malacko AR, Lioubin MN, Purchio AF. Site-directed mutagenesis of cysteine residues in the pro region of the transforming growth factor beta 1 precursor: expression and characterization of mutant proteins. J Biol Chem. 1989;264:13660–13664.
Frustaci A, Chimenti C, Bellocci F, Morgante E, Russo MA, Maseri A. Histological substrate of atrial biopsies in patients with lone atrial fibrillation. Circulation. 1997;96:1180–1184.
Janse MJ. Why does atrial fibrillation occur? Eur Heart J. 1997;18(suppl C):C12–C18.
Koh GY, Kim SJ, Klug MG, Park K, Soonpaa MH, Field LJ. Targeted expression of transforming growth factor-beta 1 in intracardiac grafts promotes vascular endothelial cell DNA synthesis. J Clin Invest. 1995;95:114–121.