Introduction
Vasculopathy is a key pathologic feature of systemic sclerosis (SSc) and leads to important clinical complications including pulmonary arterial hypertension (PAH), scleroderma renal crisis (SRC), and severe Raynaud phenomenon with digital ischemia and infarction. In this study, we explored systemic vasculopathy and cardiovascular abnormalities in a transforming growth factor-beta (TGF-β)-dependent transgenic mouse model that has been previously shown to replicate the skin and lung fibrosis of SSc.
Although many previous studies highlighted microvascular abnormalities in SSc [
1], a growing body of evidence exists for structural and functional abnormalities in the macrovascular circulation. Altered large vessel vasoreactivity and abnormal biomechanical properties have been described, including vessel stiffness and elasticity of the aorta and carotid arteries, and impaired flow-mediated dilatation in brachial arteries [
2‐
6]. Although arterial stiffness is usually considered to result in hypertension and an increased propensity to atherosclerosis and aortic aneurysm, none of these is a prevalent feature in SSc [
7]. By analogy, TGF-β overactivity is implicated in the pathogenesis of hypertensive arteriosclerosis, SSc, and some inherited vascular diseases that affect aortic structure and function, including Marfan syndrome and Loeys-Dietz syndrome [
8‐
11].
We previously described a novel genetically determined transgenic mouse strain in which ligand-dependent activation of TGF-β signaling occurs selectively in fibroblasts (Tβ RIIΔk-fib). Expression of this kinase-deficient type II TGF-β receptor at low levels facilitates activation of the endogenous type I TGF-β receptor, at least in part by increasing levels of wild-type Tβ RII. Downstream consequences include upregulation of TGF-β1 and other gene products that promote TGF-β activity or activate the latent TGF-β complex. This results in net activation of TGF-β signaling. However, in response to TGF-β1, significant elevation of transgene expression is found. Higher level transgene expression is inhibitory and blocks signaling. Thus, for transcripts upregulated at early time points by TGF-β1, a transient response occurs in transgenic cells, but for transcripts induced at 12 hours when the transgene is also upregulated, suppression is observed. High level transgene expression does not suppress the fibroblast-specific promoter completely, suggesting that other TGF-β-independent pathways also govern the activity of this lineage-specific construct. We have described this as a model of balanced TGF-β upregulation occurring selectively in fibroblasts [
12‐
15].
In the present study, we explored the potential link between TGF-β overactivity and systemic cardiovascular features of SSc. Our results show upregulation of TGF-β signalling pathways and vessel wall fibrosis in the systemic arterial circulation, altered vasoreactivity, and a TGF-β- activated smooth muscle cell phenotype with additional perturbation of endothelin axis signaling. Our work provides support for altered TGF-β activity playing a pivotal role in vasculopathy in this strain and in SSc.
Materials and methods
Generation of transgenic mice
The generation and characterization of Tβ RIIΔk-fib transgenic mice were described previously [
15]. All animal procedures were conducted in compliance with institutional and national guidelines and with ethics committee approval. Neonatal pups were genotyped by PCR analysis of genomic DNA extracted from tail-biopsy specimens, by using primers specific for the β-galactosidase reporter gene (5'-CGGATAAACGGAACTGGAAA-3' and 5'-TAATCACGACTCGCTGTATC-3') (Sigma-Genosys, Haverhill, UK).
Histologic analysis
Thoracic aortic and cardiac tissue from sacrificed adult transgenic and littermate sex-matched wild-type mice were dissected and immersed in 10% normal saline or were snap-frozen in liquid nitrogen. Formalin-fixed, paraffin-embedded composite 3 μm sections were mounted onto poly-
L-lysine-coated slides (VWR, UK) and stained with H&E, picrosirius red, Elastin van Gieson (EVG), Masson trichrome, and for immunohistochemistry according to standard protocols [
16]. Primary antibodies were as follows: CD34 (Abcam, Cambridge, UK); TGF-β1 and pSmad2/3 (Santa Cruz Biotechnology, Santa Cruz, CA); LAP(TGFβ1) (R&D Biosystems, Minneapolis, MN); and α-SMA (Sigma-Aldrich, St. Louis, MO). Mounted sections were viewed with an Axioskop Mot Plus microscope by using Axiovision software (Zeiss, Westlar, Germany). Vessel measurements were quantified using the same software.
Vascular smooth muscle cell culture
Aortae were dissected, the adventitia stripped, and the vessel opened longitudinally. After collagenase digestion (1 mg/ml) for 10 minutes at 37°C applied to the endothelial surface, the remaining smooth muscle cells were grown by explant culture in standard conditions [
17]. Immunostaining revealed >99% α-SMA positivity at day 14. Experiments were performed at passages 3 to 4. Cells were incubated for 24 hours in serum-free medium before agonist stimulation with TGF-β1 (4 ng/ml) or endothelin-1 (ET-1, 100 nmol/L). Cells were harvested using Buffer RLT (Qiagen, Crawley, UK) containing 10 μl/ml of 14.3 mmol/L β-ME or Laemmli blue buffer and stored at -70°C, or seeded into chamber slides for immunostaining.
Immunostaining of vascular smooth muscle cells
Seeded cells were fixed with methanol at -20°C for 15 seconds or paraformaldehyde for 15 minutes and rinsed in PBS. After serum blocking, the cells were stained for 1 hour at room temperature with α-SMA, anti-smoothelin, anti-SM22α, or anti-β-galactosidase (Abcam), washed in PBS, and then incubated with the appropriate secondary antibody (Vector Laboratories) in PBS for 30 minutes. The slides were washed, mounted with Vectashield mounting medium containing DAPI (Vector Laboratories), and examined with an Axioskop Z fluorescence microscope (Zeiss).
Reporter gene assay
The chemiluminescent β-galactosidase assay Galactolight Plus (Applied Biosystems, Foster City, CA) was used according to the manufacturer's instructions. In brief, equivalent numbers of vSMCs and fibroblasts in 96-well plates were lysed using the proprietary lysis buffer and incubated for 10 minutes. Then 10 μl was mixed with 70 μl reaction buffer and incubated for 1 hour; 100 μl of accelerator II was added automatically, and the luminescence was measured after 2 seconds using the Mithras LB 940 luminometer (Berthold, Wildbad, Germany). Assays were performed in triplicate.
Assay of fibrillar collagen content
Newly synthesized acid-soluble collagens (types I to IV) from the heart or aorta were quantified by using the Sircol colorimetric assay (Biodye Science, Newtownabbey, UK) according to the manufacturer's instructions and analyzed using the Mithras LB 940 plate reader. Collagen concentrations were expressed as milligrams per milliliter. Data are expressed as mean ± SEM. Statistical comparisons were made by using Student's t test.
Isometric tension measurement in isolated aortic rings
Mice thoracic descending aortae were washed in fresh Krebs buffer (119 mmol/L NaCl, 4.7 mmol/L KCl, 1.2 mmol/L MgSO4, 1.2 mmol/L KH2PO4, 11 mmol/L glucose, 25 mmol/L NaHCO3, 2.5 mmol/L CaCl2), and the loose connective tissue removed. Aortae were cut into paired 2- to 3-mm wide rings, which were mounted on two hooks in a 7 ml organ bath containing Krebs buffer at 37°C, continuously oxygenated with 95% O2/5% CO2. Isometric tension was measured with force-displacement transducers (Grass Instruments, Quincy, MA), and digitized using a multichannel recording system (Linton Instrumentation, Diss, Norfolk, UK) with MP100 acquisition unit and AcqKnowledge software (Biopac Systems, Goleta, CA). A resting tension of 500-550 mg was applied to the rings, which were then allowed to equilibrate for 60 minutes. In this period, tissues were washed out with Krebs buffer, and the applied tension readjusted at 15-minute intervals.
After the equilibration period, rings were contracted with cumulative doses of potassium chloride (KCl; 30 mmol/L and 80 mmol/L) until a stable contraction plateau was reached. Contractile responses were measured by recording changes in tension (milligrams). After washout, the tissues were allowed to reequilibrate for 30 minutes, and contractile dose-response curves were constructed using cumulative doses of phenylephrine (PE; 10-9 to 10-4 mol/L) and a stable analogue of thromboxane A2 (U46619; 10-10 to 10-4 mol/L) or ET-1 (10-11 to 10-7 mmol/L) with washout and equilibration after each dose response curve. In the relevant experiments, tissues were pretreated for 20 minutes with 2 mmol/L bosentan (a dual endothelin receptor antagonist) before contractile responses to ET-1 were measured. Data are expressed as mean ± SEM. A value of P < 0.05 was considered significant.
Quantitative RT-PCR
Total RNA was extracted by using the RNeasy minikit (Qiagen) according to the manufacturer's instructions and quantified using the Nanodrop ND-8000 spectrophotometer (Thermo-Scientific, Wilmington, DE). The minimum 260:280 ratio was 1.90. RNA integrity numbers ranged from 8.8 to 10, measured on an Agilent 2100 Bioanalyzer (Agilent Technologies UK Limited, Stockport, UK); 600 ng of RNA was reverse transcribed using the Quantitect reverse transcription kit (Qiagen) and diluted fivefold with tRNA, 0.2 μg/ml. The real-time quantitative RT-PCR used 2 μl RNA in a 10 μl reaction volume by using Sensimix NoRef in a SYBR green-based assay (Quantace, London, UK) on a Rotorgene-6000 (Corbett Life Sciences, Sydney, Australia) under the following conditions: 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds, 57°C for 10 seconds, and 72°C for 5 seconds. Specific products and absence of primer dimers were confirmed by melt curve analysis. Copy numbers and assay efficiencies were derived from known copy number standard curves. Four stable reference genes: succinate dehydrogenase complex, subunit A (Sdha); ribosomal protein L13 (Rpl13); β-actin (ActB); and ubiquitin C (Ubc) were identified by using geNorm, and copy numbers were corrected using the computed normalization factor [
18]. Primer sequences, written 5'-3', are referenced where appropriate, assay efficiency and R
2 follow: Sdha fwd TGTTCAGTTCCACCCCACA, rev TCTCCACGACACCCTTCTGT, 0.83, 0.991; Rpl13 fwd CAGTGAGATACCACACCAAGGTC, rev GTGCGAGCCACTTTCTTGT, 1.04, 0.998; ActB fwd CTAAGGCCAACCGTGAAAAG, rev ACCAGAGGCATACAGGGACA, 0.83, 0.999; Ubc fwd GAGTTCCGTCTGCTGTGTGA, rev TCACAAAGATCTGCATCGTCA, 0.93, 0.999; Pai-1 [
19], 0.89, 1.000; mCol1a1 [
19], 0.80, 0.993; Ctgf [
19] 0.91, 0.998; Smtn (smoothelin) fwd CGAGGAGGCTGCAACTTTA, rev CTGCGCCATTAGCTGCTT, 0.96, 0.999, ETRA [
20] 0.95, 0.999, ETRB [
20], 0.94, 0.999, ET-1 fwd ACTTCTGCCACCTGGACATC, rev AGTTCTTTTCCTGCTTGGCA, 0.9, 0.999; Tagln fwd GATGGAACAGGTGGCTCAAT, rev TTCCATCGTTTTTGGTCACA, 0.94, 0.999.
Floating collagen gel cultures
Experiments were performed as described previously [
21]. In brief, 24-well tissue culture plates were precoated with 2.5% bovine serum albumin (BSA). Trypsinized smooth muscle cells were suspended in Molecular, Cellular, and Developmental Biology (MCDB) 131 medium (Invitrogen, Paisley, UK) and mixed with collagen solution (one part of 0.2 mol/L
N-2-hydroxyethylpiperazine-
N'-2-ethanesulfonic acid [HEPES], pH 8.0; four parts collagen [Vitrogen-100, 3 mg/ml, Cohesion Technologies, Palo Alto, CA], and five parts of 2× MCDB) yielding a final concentration of 80,000 cells/ml and 1.2 mg/ml collagen. Collagen/cell suspension (1 ml) was added to each well. After polymerization, gels were detached from wells by adding 1 ml of medium with or without TGF-β1 (4 ng/ml). Contraction of the gel was quantified by loss of gel weight and decrease in gel diameter over a 24-hour period. Comparison of collagen gel contraction was performed by using Student's
t test. A value of
P < 0.05 was considered statistically significant.
Discussion
In this study, we examined the systemic vasculature in a mouse model of SSc in which the primary defect is fibroblast-specific perturbation of TGF-β signaling. We defined, for the first time in this strain, a structural vasculopathy with adventitial fibrosis and smooth muscle attenuation in the thoracic aorta and further demonstrated altered vasoreactivity in isolated vessel preparations in vitro. Smooth muscle cell cultures show upregulation of TGF-β- dependent genes, and cardiac fibrosis is evident. Our work complements earlier studies of skin and lung fibrosis in this transgenic mouse strain.
Previous studies of cultured cells derived from this transgenic mouse strain have focused on the properties of fibroblasts [
12]. Exploration of the biochemical and functional properties of vSMCs provides important insight into the potential pathogenic mechanisms of vascular fibrosis. The lineage-specific nature of transgene expression precludes an intrinsic perturbation of TGF-β signaling in vSMCs, as they do not express the nonsignaling type II TGF-β receptor, confirmed in Figure
3a and
3b. This explains the greater responsiveness for cardinal TGF-β-regulated transcripts that we observe in vSMCs compared with dermal fibroblasts [
12]. This is consistent with balanced upregulation of TGF-β signaling in fibroblasts
in vitro, whereas the activated phenotype of explanted vSMCs reflects previous
in vivo activation by extracellular TGF-β. Thus, alterations in vascular smooth muscle cell function are likely to reflect paracrine effects mediated by transgenic fibroblasts. This is concordant with the altered epithelial cell phenotype observed in the lungs of this mouse strain in our studies of lung fibrosis [
19], which also is attributed to bystander effects of fibroblast-dependent increased local levels of active TGF-β ligand. The alterations in endothelin signaling within the vSMCs of the Tβ RIIΔk-fib strain are reminiscent of those seen in SSc fibroblasts, which have low ETRA expression in the context of high ET-1 levels. Previous work confirmed the importance of functional cross-talk between TGF-β and ET-1 [
25,
26] in SSc pathogenesis.
Our findings extend and validate data from other TGF-β- dependent animal models of SSc. For example a rapidly progressive vasculopathy is described in the caveolin-1 knockout mouse, which occurs in part because of uncontrolled endothelial proliferation, alterations in vasomotor tone, and a fibrotic phenotype associated with increased signaling through the TGF-β axis [
27,
28], and second, the Tβ RI
CA Cre-ER mouse strain in which constitutive activation of the Tβ RI in fibroblasts results in fibrotic thickening of small vessels in the lung and kidney but histologically normal large vessels and heart [
29]. The heterozygous TSK-1 mouse, which carries a 30- to 40-kb genomic duplication in the
fibrillin-1 gene, has marked hyperplasia of loose connective tissue around the thoracic aorta [
30] and altered aortic hemodynamics
ex vivo suggestive of endothelial dysfunction [
31]. These models allow important investigation into the link between endothelial cell dysfunction and fibrosis but do not address the more chronic background vasculopathy that is a hallmark of SSc and may underlie susceptibility to important clinical complications, including PAH and SRC.
In this study, structural and dynamic alterations in large vessels are evident. Abnormalities in elasticity and compliance are most evident in patients with diffuse cutaneous SSc [
2,
6]. These result in a phenotype of arterial stiffness, which is usually considered to have independent predictive value for cardiovascular events. Whether SSc predisposes to increased atherosclerotic risk remains in question: some reports exist of increased propensity to peripheral vascular disease in limited cutaneous SSc, but an association of coronary artery disease with SSc has not been consistently demonstrated [
7,
32,
33]. Examination of the microvascular structure in this model in the future, particularly within the vascular beds of the lung, kidney, and dermis, is likely to provide further insight into the molecular basis of vasculopathy in fibrotic disorders such as SSc.
Potential mechanistic parallels exist between the Tβ RIIΔk-fib mouse strain and human Loeys-Dietz syndrome, in which mutations in Tβ RI and Tβ RII result in paradoxical increased expression of TGF-β- regulated proteins and signaling pathways. The fibroblast-specific nature of transgene expression is a likely explanation for the absence of greater phenotypic similarity in this mouse strain. In animal models of essential hypertension, arterial stiffness does not develop because of structural modifications of the vessel walls with redistribution of the mechanical load toward elastic materials. Alterations in the capacities of these remodeling processes may explain the spectrum of arterial disease seen in Marfan syndrome, Loeys-Dietz syndrome, SSc, and hypertension; fibrillin and TGF-β metabolism are implicated in all [
11,
34,
35].
The significance of myocardial fibrosis in the Tβ RIIΔk-fib strain is unclear. It may result from altered pulmonary and systemic hemodynamics or as a primary process from excessive TGF-β due to the genetic defect in the fibroblasts present within the myocardium. It is possible that an initial response to altered vascular dynamics results in increased fibroblast activity in the myocardium and hence higher expression of the transgene and upregulation of TGF-β. Autopsy studies have revealed evidence of myocardial interfascicular fibrosis and contraction-band necrosis in patients with SSc, and myocardial involvement is an adverse prognostic feature of this condition [
36]. The presence of increased cardiac collagen in this mouse strain strengthens its place as a useful disease model.
Limitations of this study include the challenge of directly extrapolating biochemical and functional results from a mouse model to a complex multisystem disease such as SSc. Moreover, it is challenging to separate primary effects of an alteration of fibroblast-derived TGF-β from those that are due to altered vascular smooth muscle cell properties. Differences may exist between
in vivo mechanisms and the properties of explanted cells in tissue culture or isolated organ bath preparations, although this method was selected as it provides one of the most physiologic platforms for studies of vasoreactivity
ex vivo. As expected from the published literature [
37], murine aortic rings were only weakly responsive to endothelin, in contrast to vessels from other species. Where the tension axis falls below zero in Figure
5c, suggesting vasodilation, we speculate that this relates to unopposed vasodilator effect of type B endothelin receptors. However, although consistent, this effect did not reach statistical significance. Technical limitations of this study include the need to perform studies on relatively small numbers of mice and the measurement of mRNA expression levels that may not correlate with function or activity of encoded protein.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
CPD had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. ECD-S, DJA, and CPD participated in the study design. ECD-S, AD, KK, and XS-W participated in the acquisition of data. ECD-S, AD, Khan, XS-W, and CPD participated in the analysis and interpretation of data. ECD-S, DJA, and CPD prepared the manuscript. ECD-S was responsible for statistical analysis.