Introduction
In North America, advances in perinatal care over the past decade have failed to reduce the occurrence of intrauterine growth restriction (IUGR) and in fact were paralleled by an increase in the number of babies born small.Reference Martin, Hamilton and Sutton1, Reference Raphael2 IUGR is the outcome of an adaptation to unfavorable intrauterine conditions during critical developmental windows and is a significant clinical event that has long-term consequences for the functional efficiency and structural integrity of organ systems. Although the fetal response to an insufficient substrate supply involves protection of the heart relative to other organs, subclinical systolic and diastolic dysfunction and cardiac morphological abnormalities are apparent in IUGR children and neonates.Reference Bahtiyar and Copel3, Reference Crispi, Bijnens and Figueras4 This early disadvantage appears to endure, as numerous studies report negative correlations between birth weight and risk for heart disease and heart failure in adult life. Understanding the developmental disturbances that underlie programming of cardiac dysfunction is critical in addressing this pre-birth component of heart disease.
For the study of developmental outcomes associated with IUGR, we use an ovine model whereby bolus injection of latex microspheres into the umbilical circulation restricts placental blood flow, produces abnormal flow in the umbilical artery and results in fetal hypoxemia and asymmetric IUGR.Reference Gagnon, Johnston and Murotsuki5, Reference Gagnon, Lamb and Richardson6 Placental insufficiency and associated fetal hypoxemia are the most common features of IUGR in human pregnancy.Reference Kingdom, Huppertz, Seaward and Kaufmann7 Other groups have demonstrated that placental embolization with chronic hypoxemia results in delayed cardiomyocyte differentiation in growth restricted ovine fetuses.Reference Louey, Jonker, Giraud and Thornburg8, Reference Morrison, Botting and Dyer9 These findings may in part explain the diminished myocardial performance reported in human IUGR neonatesReference Bahtiyar and Copel3, Reference Rizzo, Capponi, Cavicchioni, Vendola and Arduini10 and their later vulnerability to heart failure.Reference Barker, Osmond, Winter and Margetts11 Myocardial performance is also dependent on the fibrillar network that supports individual cardiomyocytes and aligns the myofibrils within the myocyte. Collagen fibers contribute to systolic contraction through coordination of myocyte shortening and are the primary determinants of compliance during diastole.Reference Graham and Trafford12 Indeed, disproportionate accumulation of collagen leading to fibrosis and the resultant systolic and diastolic dysfunction is an important process in the advancement of cardiovascular disease in human adults.Reference Czuriga, Paulus and Czuriga13
Collagen deposition is stimulated by hypoxia,Reference Chen, Zhang and Gan14 oxidative stress,Reference Zhao, Zhao, Chen, Ahokas and Sun15 mechanical overload,Reference Lucas, Zhang and Kaizheng16 and inflammation,Reference Kai, Kuwahara, Tokuda and Imaizumi17 all of which have been reported in pregnancies complicated by placental insufficiency. In parallel with interstitial collagen accumulation, transforming growth factor β (TGF-β1) induces the expression and activity of the extracellular matrix (ECM) degrading enzymes, matrix metalloproteinases (MMPs).Reference Ross and Tranquillo18 Therefore, with prolonged exposure to adverse stimuli, a cycle of protein synthesis and degradation leads to fibrosis, disorganization of the intricately structured matrix and subsequent systolic and diastolic dysfunction. Although progression of fibrosis in the diseased adult heart is well characterized, the response of fibrotic signaling in the developing heart to changes in the intrauterine environment, such as fetal hypoxemia, remains ill-defined. Augmented collagen deposition under conditions of placental insufficiency may contribute to the compromise in cardiac function demonstrated in human IUGR neonates. Therefore, this study examines the effect of umbilical–placental embolization leading to chronic fetal hypoxemia on myocardial collagen deposition in the late gestation ovine fetus. Additionally, oxidative stress and the pro-fibrotic TGF-β1 pathway were evaluated as possible stimuli leading to altered cardiac remodeling in the IUGR fetus.
Methods
Surgical procedure
All surgical and experimental procedures were approved by The University of Western Ontario Animal Use Subcommittee. Surgical preparation and experimental manipulations were performed as previously described.Reference Thompson, Richardson, Gagnon and Regnault19 Briefly, pregnant Western ewes between 112 and 114 days of gestation (term = 147 days) were induced [1 g thiopental sodium in solution, intravenously (IV) for induction, Abbott Laboratories Ltd., Montreal, QC] and maintained under general anesthesia (1–1.5% isoflurane in O2). Through an incision in the uterine wall, the lower body of the fetus was exteriorized and polyvinyl catheters placed in the right and left femoral arteries, a hind limb vein, and the maternal femoral vein. Subsequently, a catheter was placed in the amniotic cavity. After instrumentation, antibiotics were administered intra-operatively to the mother (IV) (0.2 g trimethoprim and 1.2 g sulfadorine, Schering Canada Inc., Pointe-Claire, QC), fetus (IV) and amniotic cavity (1 million IU penicillin G sodium, Pharmaceutical Partners of Canada, Richmond Hill, ON, Canada). The fetus was then returned to the uterus, the uterus and abdominal wall incisions were sutured in layers and catheters were exteriorized through the maternal flank.
The antibiotic regime was administered daily [mother (IV) 0.2 g trimethoprim; fetus (IV) and amniotic cavity 1 million IU penicillin G sodium] during the postoperative period (3–4 days). Daily samples of arterial blood were collected for measurement of maternal and fetal condition and catheters were flushed with heparinized saline to maintain patency. Animals were housed in individual metabolic cages with food and water available ad libitum.
Experimental design
Nineteen sheep were assigned to moderate or severe embolization or a control group. Experiments were initiated at a gestational age of 116 ± 2 days. Placental embolization was performed as previously described.Reference Thompson, Richardson, Gagnon and Regnault19 Microsphere injections were initiated at 1000 h on day 1 following a 2-h baseline period and repeated at 10-min intervals until a stable reduction of fetal arterial oxygen saturation at the desired level was maintained for at least an hour. The target level of arterial oxygen saturation was 40–50% for the moderately hypoxic group and 30–40% for the severely hypoxic group. On subsequent days embolization was performed after the 2-h baseline period if fetal arterial oxygenation rose above targeted values. Control fetuses were injected with saline only.
In vivo physiological parameters
On selected days (experimental day 1, 5, 8, 12 and 15) a blood sampling and cardiovascular monitoring regime was followed: fetal arterial and maternal venous blood samples were taken at 0900 h (baseline), 1300 h and 1600 h. Samples were analyzed for blood gases, lactate, glucose and pH using a blood gas analyzer (ABL-725, Radiometer, Copenhagen, Denmark) and corrected for fetal temperature (T = 39.2°C). Average fetal arterial oxygen saturation values for each fetus were calculated using blood gas measurements taken at 0900 h, 1300 h and 1600 h on each experimental day, excluding baseline of day 1.
Post-mortem and tissue preparation
On experimental day 15, after the final blood sample at 1600 h, ewes and fetuses were sacrificed with an overdose of barbiturate (30 mg pentobarbital sodium, Fatal-Plus; Vortech Pharmaceuticals, Dearborn, MI, USA). The fetus was weighed, sexed and dissected to obtain brain and liver weights. The right and left cardiac ventricles were separated and cut into segments that were either fast frozen in liquid nitrogen for polymerase chain reaction (PCR) and Western blot analyses or fixed in 4% paraformaldehyde. Fixed segments of the ventricular free wall were embedded in paraffin as cross-sections and cut in 5 μm sections that were baked onto positively charged glass plates by heating in a 50°C oven for 2 days.
RNA extraction and quantitative real-time PCR (qRT-PCR)
The Trizol method was used to extract total RNA from frozen right and left ventricular samples (Invitrogen Life Technologies Co., Burlington, ON, Canada). Each extracted sample was assessed for RNA integrity using 1.2% agarose electrophoresis with ethidium bromide staining. Complementary DNA was synthesized from 2 μg of purified RNA using oligo(dT) primers and the SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen Life Technologies Co.). Three primer pairs for each gene were designed using BLAST. The primer pair used in the experiment was chosen based on melting curve analysis and the presence of a single band of appropriate size in 1.8% agarose gel. Standard curves for each primer set (supplementary Table 1) were generated in order to determine optimal concentrations of input cDNA and PCR efficiency. PCR efficiencies for each primer set were 90–100%. cDNA products were used as templates for qRT-PCR for measurement of gene expression levels using the SYBR green system (Bio-Rad Laboratories, Mississauga, ON, Canada) and a Bio-Rad CFX384. Amplification was performed in triplicate at 95°C for 3 min, followed by 39 cycles at 95°C for 15 s, 59°C for 15 s and 72°C for 15 s. Fold change values calculated relative to the average mRNA for a given gene were normalized to a reference gene (S15).
Nuclear extraction
Fetal ventricle samples were homogenized with a pestle and sheered with a 25 G needle in fractionation buffer (250 mM sucrose, 20 mM HEPES, 10 mM KCl, 1.5 mM MgCl2, 1 mM EDTA, 1 mM EGTA) containing 1 mM DTT and protease and phosphatase inhibitor cocktail (Thermo Scientific, Nepean, ON, Canada). Centrifugation at 300 g was used to remove the debris. Next, the supernatant was centrifuged at 1500 g in order to isolate nuclear proteins which were subsequently washed with fractionation buffer and then re-suspended in RIPA lysis buffer containing 10% glycerol and 0.1% SDS. The supernatant and nuclear extraction were analyzed concurrently with Western blot for the presence of histones using a monoclonal mouse anti-histone antibody (Upstate Cell Signalling Solutions, Lake Placid, NY, USA) at a dilution of 1:1000, in order to confirm the efficiency of nuclear extraction. The supernatant was retained for cytoplasmic proteins. Protein concentration of the samples was determined using the BCA protein assay (Thermo Scientific).
Western blot analysis
Aliquots of protein extract were added to NuPAGE lithium dodecyl sulfate (LDS) sample buffer and sample reducing agent (Invitrogen Life Technologies Co.) heated for 10 min at 80°C. Ten micrograms of protein from the nuclear extracts and 30 μg of protein from the supernatant were loaded on NuPAGE 4–12% Bis-Tris polyacrylamide gels (Invitrogen Life Technologies Co.). Proteins were separated by electrophoresis using MES running buffer and electroblotted onto PVDF immobilon transfer membranes (Millipore, Etobikoke, ON, Canada) at 100 V for 2 h. Following transfer, membranes were thoroughly washed and then blocked for 1 h with 5% albumin in Tris-buffered saline containing 0.1% Tween (TBST). Membranes were then incubated overnight with primary antibodies at the following concentrations: rabbit polyclonal anti-TGF-β1 antibody directed to the TGF-β1 precursor (45–65 kDa) at 1:1000; rabbit polyclonal anti-SMAD2/3 antibody directed to the Smad2/3 (52, 60 kDa) at 1:1000; rabbit monoclonal anti-phospho-Smad3 (Ser423/425) directed to P-Smad-3 (52 kDa) at 1:1000; rabbit polyclonal anti-phospho-Smad2 (Ser465/467) directed to P-Smad-2 (60 kDa) at 1:1000; rabbit polyclonal anti-Smad-4 antibody directed against Smad 4 (70 kDa) at 1:2000; rabbit monoclonal anti-ERK1/2 directed to ERK1/2 (44, 42 kDa) at 1:500; rabbit monoclonal anti-phospho-p44/42 MAPK (ERK1/2) (Thr202/Tyr204) directed to ERK1/2 (44, 42 kDa) at 1:500; rabbit monoclonal anti-P38 MAPK directed to P38 MAPK (40 kDa) at 1:500; rabbit monoclonal anti-phospho-P38 MAPK directed to phospho-P38 (40 kDa) at 1:500. All antibodies were purchased from Cell Signaling (Danvers, MA, USA). The following day, membranes were washed with TBST and probed with a rabbit or mouse polyclonal secondary antibody conjugated with HRP (Jackson Immunoresearch, West Grove, PA, USA) for 1 h at room temperature. The immunoreactivity was visualized by chemiluminescence and densitometry performed. For nuclear proteins, histone 3 was used as a loading control, all other protein bands were normalized against sample specific protein bands stained with Amido Black.
Sirius red staining
After deparaffinization in xylene, slides were rehydrated by passage through a decreasing ethanol series. Collagen content was measured in cross-sections stained with 1% Sirius red F3BA (Sigma-Aldrich Canada Ltd., Oakville, ON, Canada) in a saturated aqueous solution of picric acid, for 1 h. Sirius red binds specifically to fibrillar collagens I and III.Reference Lee, Assoku and Doyle20 Since the goal of the present study was to measure total collagen volume, Sirius red-stained samples were analyzed using light microscopy rather than the polarization method which distinguishes between the collagen types through their associated variations in color and intensities of birefringence. Stained cross-sections were captured on a microscope (Leica DM RB) at ×40 magnification. The area positive for collagen was identified by color thresholding using image analysis software (Image Pro 6.0; MediaCybernetics, Bethesda, MD, USA) and expressed as a percentage of the sum of area stained. All slides were stained in duplicate and 7–10 areas per slide were analyzed. The operator was blinded to animal identification.
Nitrotyrosine staining
Nitrotyrosine immunoreactivity was assessed using immunofluorescence as a measure of oxidative stress. Sections of free ventricular wall cross-sections were deparaffinized and incubated for 8 min in Background SNIPER (Biocare Medical, Concord, CA, USA) to reduce non-specific background staining. Sections were then incubated in rabbit anti-nitrotyrosine antibody (Santa Cruz Biotechnology Inc.) at a dilution of 1:1000 overnight at 4°C. After washing in phosphate-buffered saline, slides were incubated with secondary antibody, goat anti-mouse Alexafluor 568 (1:200; Invitrogen Life Technologies Inc.) at room temperature for 30 min in a black covered humidity chamber. Replacement of the primary antibody with PBS or IgG was used as negative controls. Fluorescence VectaShield mounting medium (Vector Laboratories, Burlington, ON, Canada) was used for mounting. Slides were stained in duplicate and simultaneously to minimize variation in staining intensity. The nitotyrosine antibody was tested for specificity by Western blot. Sections were imaged on a microscope (Zeiss Z1) and captured at ×20 magnification using a camera and software for image capture and analysis (Axiovision 4.0; Carl Zeiss Microimaging LLC, Thornwood, NY, USA). The intensity of staining was measured using image analysis software (Image Pro 6.0; MediaCybernetics). All slides were stained in duplicate and 7–10 areas per slide were analyzed. The operator was blinded to animal identification.
Statistical analysis
All statistical procedures were performed using GraphPad Prism 5.0 (GraphPad Software, San Diego, CA, USA). Between group differences were assessed using one-way ANOVA with Bonferroni post-hoc tests for pair-wise comparisons, or Kruskal-Wallis with Dunns post-hoc test when data were determined to be non-parametric. All data are presented as mean±s.e.m. Statistical significance was assumed when P values were < 0.05.
Results
Of the 19 animals that originally began the experimental procedures, six animals in each of the control and severe embolization groups and four animals in the moderate embolization group were included in data analyses. Fetuses from two ewes that developed mastitis during the experiment and one fetus that was severely hypoxic during and following surgery were excluded from the study.
Morphometrics and in vivo parameters
Organ weight and fetal size have been reported separately.Reference Thompson, Richardson, Gagnon and Regnault19 Fetuses subjected to severe hypoxia, through repeated microsphere injection, were asymmetrically growth restricted as reflected by reduced fetal body weight (P < 0.05) and increased brain-to-liver ratio (P < 0.05), relative to control. Moderately growth restricted fetuses displayed a reduction in body weight that was not statistically significant and no difference in brain-to-liver ratio, compared with control. The weight of the heart relative to body weight was not different between the groups.
Detailed analyses of the results for blood pressure, heart rate, blood gas values, pH, hemoglobin, glucose and lactate concentrations have been reported previously.Reference Thompson, Richardson, Gagnon and Regnault19 Hypoxia of a moderate or severe degree for the two experimental groups was not accompanied by cumulative metabolic acidosis or lactic acid accumulation. Fetal arterial oxygen saturation averaged 51.3 ± 0.8% for the control group, 44.1 ± 1.1% for the moderate group and 35.6 ± 1.3% for the severe group (control v. moderate, P < 0.01; control v. severe, P < 0.001; moderate v. severe, P < 0.001). Baseline values of MAP and FHR were not different between groups or within groups over the 15-day study period (included in analysis are days 1, 5, 8, 12, 15).
Procollagen I and III mRNA expression and fibrillar collagen content
Fibrillar collagens are the primary determinants of ventricular compliance; the main fibrillar collagens in the heart are collagen I and III.Reference Graham and Trafford12, Reference Czuriga, Paulus and Czuriga13 mRNA levels of collagen gene products, procollagen I and procollagen III, were increased by two-fold (P < 0.05) and three-fold (P < 0.05), respectively, in the right ventricle of severely hypoxic fetuses relative to control (Fig. 1a). In the left ventricle procollagen I mRNA levels were increased in the left ventricle of severely hypoxic animals (P < 0.05), while no differences were found with respect to procollagen III mRNA (Fig. 1b). In cross-sections of the right free ventricular wall, the percent area stained with picric-acid Sirius red, as a measure of total fibrillar collagen, was increased two-fold in the severe group, compared with both moderate (P < 0.01) and control groups (P < 0.05) (Fig. 1d). Fibrillar collagen content in the left ventricle was not different across groups.
Procollagen IαI mRNA levels in the right and left ventricle of control, moderately hypoxic and severely hypoxic fetuses (a). Procollagen III mRNA levels in the right and left ventricle of control, moderately hypoxic and severely hypoxic fetal sheep (b). The % staining of fibrillar collagen fibers as measured by picric-acid Sirius red in cross-sections of the right ventricle of each group (c). Shown is a representative right ventricular cross-section stained with picric-acid Sirius red from the control, moderate and severe group (d). *P < 0.05 control v. hypoxic group. ††P < 0.01 moderate v. severe group.
TGF-β1, MTI-MMP and MMP-2 mRNA expression
TGF-β1 is an important mediator of increased collagen synthesis and also induces both the expression and activity of MMP-2. Activation of MMP-2 is accomplished through proteolysis by the stromelysin, membrane-type MMP (MTI-MMP) and studies have demonstrated that the activity of MMP-2 is tightly correlated with the expression of MTI-MMP.Reference Cowell, Knauper and Stewart21 Therefore, we used the expression levels of MTI-MMP as an indicator of MMP-2 activity. In the right ventricle, severe hypoxia over 15 days resulted in a two-fold (P < 0.05) increase in mRNA levels of the pro-fibrotic cytokine TGF-β1 and a 1.5-fold increase in protein levels of the precursor to TGF-β (Fig. 2a). In the left ventricle, the mean values of TGF-β1 mRNA levels were increased in severe groups and TGF-β protein levels were also increased by 1.4-fold in the severe group relative to control (Fig. 2b). MTI-MMP mRNA levels were increased in the right ventricle of the severely hypoxic group by two-fold (P < 0.05), relative to control, while its primary substrate, MMP-2 showed a similar trend (P = 0.06) (Fig. 3a).
TGF-β1 mRNA levels in the right and left ventricle of control, moderately hypoxic and severely hypoxic fetal sheep (a). Protein levels of the precursor to TGF-β in the right ventricle (control v. severe, P = 0.054) and left ventricle (b, c). Shown are representative blots of the intracellular signaling components of TGF-β, including ERK and SMAD-4 (c). *P < 0.05 hypoxic group v. control.
MMP-2 and MTI-MMP mRNA in the right ventricle (a) and in the left ventricle (b). *P < 0.05 control v. hypoxic group.
SMAD2/3/4, ERK1/2 and P38
TGF-β1 primarily signals through the SMAD proteins. Upon binding of TGF-β1 to the type I and type II receptors, SMAD2/3 becomes phosphorylated and forms a complex with SMAD 4, which then translocates to the nucleus to alter gene transcription.Reference Nakerakanti, Bujor and Trojanowska22 Additionally, the mitogen-activated protein kinases (MAPK) including P38 and ERK1/2, transmit signals generated by cytokines, growth factors and environmental stresses and are involved in TGF-β1-induced collagen synthesis independent of SMADs.Reference Nakerakanti, Bujor and Trojanowska22 Across the three study groups, there were no differences in protein levels of SMAD 4, total SMAD 2/3, phosphorylated SMAD-2, phosphorylated SMAD-3 and in the ratio of total SMAD2/3:phosphorylated SMAD-2 or SMAD-3. As well, total ERK, phosphorylated ERK, P38 and phosphorylated P38 and their respective ratios, were not different between the three study groups. P-ERK, ERK and SMAD-4 are shown in Figure 2c.
Nitrotyrosine staining
Nitrotyrosine is generated by reactive nitrogen species and is a common marker of oxidative stress. The average intensity of nitotyrosine staining in the right ventricle was 22.3 ± 1.1 AU for the control group, 24.8 ± 2.0 AU for the moderate group and 34.3 ± 6.24 AU for the severe group. While four of the six animals in the severe group displayed staining similar to the moderate and control group, the two most severely hypoxic animals with oxygen saturations of 33.2% and 36.8%, had staining intensities of 40.9 and 54.7 AU, respectively, which are approximately two-fold the mean of all animals excluding these two animals.
Discussion
This study demonstrated increased collagen deposition in the heart of near-term ovine fetuses in response to chronic (15 days) hypoxia in the right ventricle. Chronic fetal hypoxemia is a common complication of human pregnancyReference Ghidini23 and was produced in the present study through umbilical–placental embolization that recapitulates the severe type of human placental insufficiency characterized by reduced villous branching, avascular villi and thickened exchange barrier.Reference Gagnon, Johnston and Murotsuki5, Reference Krebs, Macara, Leiser and Bowman24, Reference Macara, Kingdom and Kohnen25 Advantages of this animal model are the ability to monitor and control the degree of fetal hypoxemia, in vivo measurement of maternal and fetal physiological parameters under steady-state conditions and the credibility in extrapolation to human pregnancy. With respect to the latter, the developmental time course of the major organ systems is comparable between sheep and humans. For instance, cardiac development in the ovine fetus parallels that of the human in temporal aspects of blood pressure elevation and maturational events such as cellular differentiation.Reference Adler and Costabel26, Reference Burrell, Boyn and Kumarasamy27
IUGR has been associated with cardiac dysfunction in the fetal, neonatal and postnatal periodsReference Bahtiyar and Copel3, Reference Crispi, Bijnens and Figueras4, Reference Rizzo, Capponi, Cavicchioni, Vendola and Arduini10 and with a propensity toward heart failure in adulthood.Reference Barker, Osmond, Winter and Margetts11 Results of the present study suggest that this cardiac dysfunction may be in part attributable to increased collagen accumulation in utero. Fibrillar collagen content is the major determinant of ventricular compliance during diastole and contributes to systolic function through coordination of myocyte shortening.Reference Graham and Trafford12 Evidence of decreased ventricular compliance in human IUGR fetuses has been provided by Doppler measurement of reduced deceleration times at the atrioventricular valves and reduced ratio between early and late ventricular filling velocities (E/A ratio) in both the left and right ventricle.Reference Hernandez-Andrade, Crispi and Benavides-Serralde28 Further, longitudinal studies have demonstrated that this diastolic dysfunction occurs progressively with advancing gestation and appears first in the right ventricle, as early as mid-gestation.Reference Benavides-Serralde, Scheier and Cruz-Martinez29, Reference Takumi30 Thus, it appears that the present study captured this early effect on the right ventricle, as fibrillar collagen content and collagen gene expression were increased in the right ventricle at day 15 of chronic hypoxia. Increased afterload due to hypoxic-induced peripheral vasoconstriction of the fetal vessels and increased placental resistance, which occur during both umbilical–placental embolization and human placental insufficiency, are predominately imposed on the right ventricle.Reference Benavides-Serralde, Scheier and Cruz-Martinez29, Reference Arduini, Rizzo and Romanini31 This is because in the fetal circulation, the right ventricle is the primary supplier to the descending aorta and placenta. In contrast, vasodilation of the coronary and cerebral circulations mitigates the afterload effect on the left ventricle. Although our findings reveal a fibrotic response in only the right ventricle, likely reflective of this early hemodynamic response to hypoxia, they do not preclude the possibility of altered collagen accumulation in the left ventricle with an increased duration of hypoxia. In fact, protein levels of TGF-β which is a key mediator of collagen synthesis were increased at 15 days of hypoxia in the left ventricle of severely hypoxic animals, more so than in the right ventricle. Certainly, this study is limited in that 15 days of hypoxia is acute relative to human placental insufficiency. Nevertheless, the data provide evidence that collagen synthesis and deposition are increased under conditions of placental insufficiency and associated hypoxia.
As potential mediators of the increase in collagen deposition observed, we explored the fibrotic pathways involved in collagen gene expression and synthesis. The pro-fibrotic cytokine, TGF-β1 is a key player in ventricular hypertrophy in response to increased loading conditions, oxidative stress and neurohormonal signalingReference Lucas, Zhang and Kaizheng16, Reference Kai, Kuwahara, Tokuda and Imaizumi17 and is an important regulator of heart development through its effects on the ECM and cellular differentiation. TGF-β1 induces the expression of the precursor proteins for collagen I and III and regulates the expression of proteases such as the basement membrane-degrading gelatinase, MMP-2 and its activator, MTI-MMP.Reference Ross and Tranquillo18, Reference Cowell, Knauper and Stewart21 Once exposure to adverse stimuli becomes chronic, continual TGF-β1 and MMP-2 activation in the myocardium leads to excess collagen accumulation, disorganization of the ECM network and associated functional decline. The increase in collagen I and III mRNA levels in the right ventricle observed in the present study was accompanied by increases in TGF-β1 mRNA levels as well as an increase in mRNA levels of MTI-MMP. This increase in MTI-MMP suggests that the activity of MMP-2 is concomitantly increased as previous studies have demonstrated the activity of MMP-2 to be closely parallel with the gene expression of MTI-MMP.Reference Zahradka, Harding and Litchie32
The primary effectors of TGF-β signaling are the SMAD proteins. Upon phosphorylation of the TGF-β-B receptor, the receptor activated SMADs-2 and 3 become phosphorylated and dimerize with SMAD-4.Reference Nakerakanti, Bujor and Trojanowska22 This dimer then translocates to the nucleus to initiate gene transcription. No apparent changes in protein levels of SMAD 2, 3 or 4 were observed in the hearts of our severely hypoxic fetuses. Similarly, no changes were found in protein levels of P38 and ERK1/2, which are components of the MAPK intracellular transduction pathways involved in TGF-β signaling, the former through induction of SMAD and the latter independent of SMAD.Reference Nakerakanti, Bujor and Trojanowska22 Therefore, activation of the intracellular mediators of TGF-β-induced collagen deposition likely occurred at an earlier time point. A limitation of this study is the single time-point at which measurements are taken, that is after 15 days of hypoxia, which make it impossible to identify all the molecular events that led to the increase in cardiac collagen content. However, it is interesting that TGF-β1 and procollagen gene expression in the right ventricle remain elevated at 15 days of hypoxia without increases in the intracellular mediators, suggesting that this fibrotic pathway may have become dysregulated. Further, increased protein levels of TGF-β1 in the left ventricle suggest that a fibrotic response leading to a notable increase in collagen content may have later evolved in the left ventricle with prolongation of hypoxia.
Redox signals serve critical regulatory functions in normal gestation and thus modulation of these pathways is a primary means by which intrauterine stress exerts its influence on fetal development. In the present study we used nitrotyrosine staining as a marker of oxidative stress in the fetal heart. Nitration of tyrosine residues is induced by the potent oxidant species, peroxynitrite, which in turn is a product of the interaction between nitric oxide and superoxide.Reference Pacher, Beckman and Liaudet33 Increased nitrotyrosine staining was present in the heart of the two most severely hypoxic fetuses suggesting that oxidative stress was induced under this degree of hypoxia. ROS generated during oxidative stress are known to stimulate collagen accumulation through activation of TGF-β and MMP-2.Reference Shah, McDermott and Grieve34 Evidence for a role of ROS in collagen deposition has been demonstrated in fetal guinea pig fetuses made hypoxic through maternal hypoxia.Reference Evans, Liu, Pinkas and Thompson35
Conclusions
Collagen gene expression and accumulation were increased in the right ventricle of chronically hypoxic ovine fetuses. These changes occurred with increases in TGF-β1 along with its transcriptional target, MTI-MMP, which is a primary activator of MMP-2. Our data demonstrate that the response of ECM signaling pathways in the developing heart is dependent on the severity of hypoxia, as none of our outcomes were altered under conditions of moderate hypoxia. It is important to note, that both moderate and severe hypoxia occurred without metabolic acidosis or lactic acid accumulation and thus these ovine fetuses represent human fetuses likely to survive placental insufficiency. Overall, these findings provide evidence that increased collagen deposition occurs in chronically hypoxic fetuses and suggest this to be a contributor to the compromised systolic and diastolic dysfunction reported in human IUGR fetuses and neonates. Further, since human IUGR offspring are at increased risk for heart failure and rodent IUGR offspring exhibit cardiac fibrosis in adulthood, the permanence of changes in gene expression found in the present study and how they may translate to long-term compromise in molecular and structural homeostasis, warrants further investigation.
Acknowledgements
The authors thank Brad Matushewski for his technical expertise required for performance of surgeries. The daily management of sheep experiments was accomplished through team work and the authors would like to acknowledge the other members of this team including graduate students, Andrew Prout, Jeremy McCallum and Ashley Keen as well as the post-doctoral fellow, Dr Martin Frasch. They also acknowledge Karen Nygard and Dr Lin Zhao for their technical expertise and guidance in performing the histological and molecular techniques, respectively.
Financial Support
This work was supported by grants from the Canadian Institute of Health Research (MOP-15265); and the Children's Health Research Institute. J.A.T. was a recipient of the Ontario Graduate Scholarship and the Ontario and the Ontario Graduate Scholarship in Science and Technology.
Conflicts of Interest
None.
Ethical Standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national guides on the care and use of laboratory animals (Canadian Council on Animal Care Regulations) and the methods were approved by the institutional committee (The University of Western Ontario Animal Use Subcommittee, 2005-061-09).
Author Contributions
Jennifer Thompson assisted in all surgeries, performed both in vivo and post-mortem data collection, conducted the experimental paradigm along with a team, performed the molecular analyses, interpreted the data and wrote the manuscript. Bryan Richardson, along with Robert Gagnon and Timothy Regnault, conceived the experimental paradigm. Karolina Piorkowska performed nitrotyrosine staining and image capturing and edited the manuscript. Timothy Regnault provided guidance in molecular analyses and data interpretation and edited the manuscript. All authors have approved the manuscript.
