SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Objective

Identification of isolated congenital heart block (CHB) predicts, with near certainty, the presence of maternal anti-SSA/Ro antibodies; however, the 2% incidence of CHB in first offspring of anti-SSA/Ro+ mothers, 20% recurrence in subsequent pregnancies, and discordance in identical twins suggest that an environmental factor amplifies the effect of the antibody. Accordingly, this study was carried out to explore the hypothesis that hypoxia potentiates a profibrosing phenotype of the fetal cardiac fibroblast.

Methods

Evidence of an effect of hypoxia was sought by immunohistologic evaluation of CHB-affected fetal heart tissue and by determination of erythropoietin levels in cord blood. The in vitro effect of hypoxia on gene expression and phenotype in fibroblasts derived from fetal hearts and lungs was investigated by Affymetrix arrays, quantitative polymerase chain reaction (PCR), immunofluorescence, and immunoblotting.

Results

In vivo hypoxic exposure was supported by the prominent intracellular fibroblast expression of hypoxia-inducible factor 1α in conduction tissue from 2 fetuses in whom CHB led to death. The possibility that hypoxia was sustained was suggested by significantly elevated erythropoietin levels in cord blood from CHB-affected, as compared with unaffected, anti-SSA/Ro–exposed neonates. In vitro exposure of cardiac fibroblasts to hypoxia resulted in transdifferentiation to myofibroblasts (a scarring phenotype), as demonstrated on immunoblots and immunofluorescence by increased expression of smooth muscle actin (SMA), an effect not seen in lung fibroblasts. Hypoxia-exposed cardiac fibroblasts expressed adrenomedullin at 4-fold increased levels, as determined by Affymetrix array, quantitative PCR, and immunofluorescence, thus focusing attention on cAMP as a modulator of fibrosis. MDL12,330A, an adenylate cyclase inhibitor that lowers the levels of cAMP, increased expression of fibrosis-related proteins (mammalian target of rapamycin, SMA, plasminogen activator inhibitor type 1, and type I collagen), while the cAMP activator forskolin attenuated transforming growth factor β–elicited fibrosing end points in the cardiac fibroblasts.

Conclusion

These findings provide evidence that hypoxia may amplify the injurious effects of anti-SSA/Ro antibodies. Modulation of cAMP may be a key component in the scarring phenotype. Further assessment of the susceptibility of cardiac fibroblasts to cAMP modulation offers a new research direction in CHB.

Congenital heart block (CHB) which occurs without structural abnormalities and is detectable in the second trimester is almost universally associated with maternal IgG autoantibodies reactive with the intracellular soluble ribonucleoproteins 48-kd SSB/La, 52-kd SSA/Ro, and 60-kd SSA/Ro (1). Although this remarkable association suggests an inherent pathogenicity of the antibody, only 2% of neonates born to mothers with these candidate autoantibodies have CHB (2). In fetuses exposed to maternal anti-SSA/Ro and anti-SSB/La antibodies, the pathway to clinical effect may be variable; in most fetuses, the effects are kept in check (normal sinus rhythm), while in others they are subclinical (first-degree block) or, in a very few, are fully executed (advanced block with complete fibrosis of the atrioventricular [AV] node and cardiomyopathy secondary to endocardial fibroelastosis). In addition, the recurrence rates of CHB in subsequent pregnancies approach 20%, not 100% (1). Fetal genetic factors may play a contributory role, but are not fully causative, since identical twins are more often discordant than concordant for the disease (for review, see ref.1). These disparities imply that maternal antibodies are a necessary component of the pathway, and that fetal factors are perhaps additive; however, even when acting in concert, these elements are insufficient to cause damage. The rarity of fully executed CHB may be the consequence of a triple hit: maternal antibodies, fetal factors, and the in utero environment.

Based on the immunohistologic findings of fatal CHB and in vitro coculturing of fetal cardiocytes, macrophages, and fibroblasts in our laboratory, we have proposed a pathologic cascade to cardiac injury (3–5). The scenario initiates with exaggerated apoptosis (observed in the septal tissue of all hearts with CHB studied to date). Apoptosis results in surface translocation of the normally sequestered intracellular target autoantigens, which accounts for the accessibility of these autoantigens to circulating maternal autoantibodies. Secretion of profibrosing cytokines, such as transforming growth factor β (TGFβ), by macrophages that have phagocytosed the opsonized apoptotic cardiocytes promotes the transdifferentiation of the cardiac fibroblasts to a scarring phenotype. The histologic hallmark of CHB is fibrosis, which is remarkable, given the prevailing dogma that fetuses heal without permanent scarring. Equally remarkable is that AV nodal replacement occurs rapidly, with bradyarrhythmia reported, in some cases, within 2 weeks after normal sinus rhythm (see ref.6 and Buyon JP: unpublished observations). Accordingly, focus on the fetal cardiac fibroblast should be a priority of translational studies in CHB.

A potential contribution by hypoxia is underscored by the fact that the fetal heart may be exposed to a relative state of chronic hypoxia, especially when compared with adult hearts (which are not affected despite circulating anti-SSA/Ro antibodies), since the coronary artery in adults, but not fetuses, is the first branch drawing fully oxygenated blood. Moreover, cAMP, a key regulator of homeostasis of extracellular matrix in fibroblasts, may be altered during hypoxic conditions (7). It is possible that cAMP is maintained at relatively high levels during normal development of the fetal heart, as was shown indirectly by a recent report indicating significantly higher levels of cAMP in cord blood compared with that obtained from adults (8).

Accordingly, the present study was initiated to experimentally address the effect of hypoxia and TGFβ on the fibrosing component of CHB. Evidence of hypoxic exposure was sought by immunohistologic evaluation of the cardiac tissue from fetuses in whom CHB led to death, and by determination of the levels of erythropoietin in the cord blood of anti-SSA/Ro–exposed CHB-affected and unaffected neonates. Fibroblasts isolated from human fetal hearts and lungs (a nonsusceptible organ) were cultured under conditions of hypoxia and environmental modulation of cAMP, and evaluated for expression of messenger RNA (mRNA) and protein related to homeostasis of extracellular fibrous matrix.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Isolation and culture, under normoxic and hypoxic conditions, of cardiac and lung fibroblasts.

Fibroblasts were isolated from human fetal heart and lung tissue, as described previously (9). Briefly, hearts and lungs were aseptically dissected from aborted human fetuses of gestational age, 16–24 weeks after elective termination of normal pregnancy by dilatation and evacuation. This was done in accordance with the guidelines of our Institutional Review Board and after obtaining consent from the mothers. For isolation of cells from the heart, the aorta was cannulated for continuous perfusion of the coronary arteries using a Langendorff preparation (9). The heart was treated with collagenase A (type III), which was recirculated for 20 minutes. The heart dissociated spontaneously, allowing cells to slowly drip and fall on a petri dish containing 0.25% trypsin and 1 mM EDTA in Hanks' balanced salt solution. Clumps of cells were dissociated and the resulting suspension was poured over a cell strainer. Cells were centrifuged and the pellet was resuspended in 20 ml of culture medium (Dulbecco's modified Eagle's medium [DMEM] supplemented with 10% fetal bovine serum (FBS), 50 units/ml penicillin, 50 units/ml streptomycin, 100 mg/ml gentamicin, 1 mM nonessential amino acid [Life Technologies, Rockville, MD], 0.1 mM essential medium vitamins [Life Technologies], 2 mM glutamine, 0.1 mM sodium pyruvate).

The cell isolate contained both cardiac myocytes and cardiac fibroblasts. Separate enriched cultures of each cell type were generated by an initial adhesion step in which 1.2 × 107 cells were plated per 75-cm2 culture flask in DMEM plus 10% FBS (for 20 minutes at 37°C). Cardiac myocytes (comprising the nonadherent cells) were centrifuged for the purpose of exclusion from these studies. Fibroblasts (adherent cells) at passages 1–3 were routinely used in these studies. The fibroblast enrichment in the cell culture (fibroblasts proliferate rapidly compared with myocytes) was observed to be higher than 90%, as assessed using a monoclonal antibody clone, IB10 (F-4771; Sigma-Aldrich, St. Louis, MO), that recognizes fibroblasts. The experiments utilized cardiac fibroblasts isolated from 11 separate donors, in addition to 3 sets in which fibroblasts from the hearts and lungs were obtained from the same donor.

To obtain pulmonary fibroblasts, explants were prepared by mincing tissue. After 3 days in DMEM plus 10% FBS, clumps of cells were evident as rings of cells surrounding the explants. These pulmonary fibroblasts were enzymatically isolated using trypsin, and the primary isolate was plated in flasks (for 20 minutes at 37°C). Fibroblasts at passages 1–3 were routinely used in these studies.

All experiments were performed at <30% confluence. Cells were cultured in an incubator under 5% CO2 atmosphere at 37°C, to maintain conditions of normoxia. Under these normoxic conditions, cells were cocultured with the adenylate cyclase inhibitor MDL12,330A (30 μM for 5 hours; Calbiochem, La Jolla, CA) or with the cAMP activator forskolin (10 μm for 5 hours; Sigma-Aldrich). In addition, under normoxic conditions, cells were cultured in the presence of TGFβ (10 ng/ml; R&D Systems, Minneapolis, MN).

For exposure to hypoxic conditions, cells were maintained as described previously (10), in a controlled atmosphere chamber (Plas-Labs, Lansing, MI) at 37°C. Oxygen tensions were regulated by a palladium catalyst, which converts environmental oxygen and hydrogen to water. Oxygen was maintained at <0.1%, and its level was measured with a calibrated Series 200 Percent oxygen analyzer (Plas-Labs); the pH remained stable throughout the experiments in hypoxic conditions.

Lactate dehydrogenase (LDH) was measured by specific enzyme-linked immunosorbent assay (ELISA), according to the manufacturer's directions (R&D Systems).

Tissue sections from fetal hearts.

Formalin-fixed paraffin-embedded sections were obtained from the hearts of 2 fetuses with fatal CHB (their clinical description and gross anatomy have been described previously [3]) and from the heart of a normal human fetus electively terminated at 23 weeks' gestation. For immunostaining, sections of the fetal heart tissue were prepared as described previously (3). Briefly, anti–mammalian target of rapamycin (anti-mTOR) (Stressgen Biotech, Victoria, British Columbia, Canada), rabbit IgG (isotype control to anti-mTOR; Southern Biotechnology, Birmingham, AL), anti–plasminogen activator inhibitor type 1 (anti–PAI-1) (Transduction Laboratories, San Diego, CA), anti–hypoxia-inducible factor 1α (anti–HIF-1α) (Upstate Biotechnology, Lake Placid, NY), anti–smooth muscle actin (anti-SMA; Sigma-Aldrich), or mouse IgG as isotype control (Accurate Chemical and Scientific, Westbury, NY) were used as the primary antibodies. Stains were visualized using anti-rabbit or anti-mouse IgG peroxidase (brown) or anti-mouse IgG alkaline phosphatase (red). Sections were counterstained prior to obtaining photographic images.

Immunofluorescence and immunoblot analyses of human fetal fibroblasts.

For immunofluorescence, human fetal fibroblasts were fixed with paraformaldehyde and processed, using standard techniques, with the same primary antibodies as described above. In addition, mouse anti-human type I collagen (Upstate Biotechnology) was used as a primary antibody. Secondary antibodies were anti-mouse IgG tetramethylrhodamine isothiocyanate or anti-rabbit IgG rhodamine (Sigma-Aldrich). The results of immunofluorescence analysis were obtained using an immunofluorescence microscope equipped with a digital camera.

For immunoblotting, human fetal fibroblast suspensions were isolated and then lysed in sodium dodecyl sulfate (SDS)–sample buffer. In addition, the primary antibody mouse anti-human tubulin (Sigma-Aldrich) was used. Immunoblotting was performed using standard techniques, with the same primary antibodies as described above.

Microarray analysis.

Replicate flasks were seeded with cardiac fibroblasts, and cells were grown until confluent. After 4 hours of hypoxia (<0.1% oxygen and 5% CO2), cells were harvested with a cell scraper. RNA was isolated using TRIzol and directly processed for preparation of complementary DNA (cDNA) and hybridization to filters. Five micrograms of RNA, isolated from control and treated cells, was used for microarray analysis. The DNA array analysis was performed at the New York University Cancer Institute genomics facility; biologic samples were processed in duplicate for each experimental condition.

The RNA quality was assessed on a Bioanalyzer 2100 (Agilent Technologies, Palo Alto, CA). For each sample, 50 ng total RNA was amplified and labeled with the 2-Cycle cDNA Synthesis and the 2-Cycle Target Labeling and Control Reagent packages (Affymetrix, Santa Clara, CA) following the manufacturer's recommendations. Biotin-labeled fragmented complementary RNA (15 μg) was hybridized to GeneChip Human Genome U133A 2.0 arrays (Affymetrix), which can identify more than 18,400 transcripts representing 14,500 human genes.

Hybridization occurred at 45°C for 16 hours in a Hybridization Oven 640 (Affymetrix). Chips were then washed and stained in a Fluidics Station 450 (Affymetrix) and scanned with a GeneChip Scanner 3000 (Affymetrix). Raw GeneChip data were normalized at the probe level using the robust multichip average algorithm (11), and further filtered using GeneSpring 7.2 (Agilent Technologies). The differentially abundant mRNA (between normoxic and hypoxic conditions) were statistically filtered using an intersection of t-test results (with P values of 0.05 considered significant) and significance analysis of microarrays (with a false discovery rate set at 5%) (12).

Real-time quantitative polymerase chain reaction (PCR) for validation of gene expression data.

Quantitative PCR was performed under standard conditions using the primers listed below. Each sample was assessed in triplicate, and a reverse transcriptase negative control was also tested to exclude any contamination from DNA amplification. The relative expression ratios were calculated using the 2math imagemethod (13). The expression level of the β2-microglobulin (B2M) gene was used to normalize for differences in input cDNA. The primers for B2M were AAAGATGAGTATGCCTGCCG and CCTCCATGATGCTGCTTACA. The primers for hypoxia-inducible protein 2 (HIG2) were TGGAGTCCCTAGAGGGCTTA and TTGGCTAGTTGGCTTCTGGT. The primers for adrenomedullin were AGGACATGAAGGGTGCCTCT and GTAGCGCTTGACTCGGATG. The primers for vascular endothelial growth factor (VEGF) were CGCAAGAAATCCCGGTATAA and AAATGCTTTCTCCGCTCTGA.

Cord blood samples and ELISA for erythropoietin.

Cord blood samples were obtained at birth from family members enrolled in the Research Registry for Neonatal Lupus (14) as well as the PR Interval and Dexamethasone Evaluation in CHB Study (15). For inclusion in the present study, a child was considered to have CHB if the following 2 criteria were met: 1) presence of heart block (first-, second-, or third-degree) documented by electrocardiogram, echocardiogram, history of pacemaker, or statement in the medical record; and 2) presence of antibodies to SSA/Ro in the maternal serum, as determined by commercial ELISA (Diamedix, Miami, FL), ELISA with recombinant proteins, and/or SDS immunoblotting to identify the fine specificity of the autoantibody response. Unaffected children were exposed to maternal antibodies as defined above, but had no evidence of heart block.

The serum concentration of erythropoietin was determined using an immunoassay (catalog no. 01630; StemCell Technologies, Vancouver, British Columbia, Canada) according to the manufacturer's instructions.

Statistical analysis.

Data were analyzed using the Mann-Whitney U test, to determine statistically significant differences in erythropoietin cord blood levels between CHB-affected and unaffected neonates. An additional analysis utilizing Fisher's exact test was carried out to compare the proportions of each group that exceeded the 90th percentile of the values in the healthy control group. P values less than 0.05 were considered significant.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

In vivo evaluation of hypoxic exposure and fibrosis.

Initial experiments were performed to address whether cardiac tissue from the 2 CHB-affected fetuses had evidence suggestive of hypoxia-induced injury. Prominent intracellular fibroblast expression of HIF-1α was seen in the region of the cardiac tissue from the 20-week-old CHB-affected heart, within areas adjacent to zones of fibrosis and calcification (Figure 1A). As previously demonstrated in the septal tissue of this 20-week-old CHB-affected heart, myofibroblasts exhibiting expression of SMA were observed at the anticipated sites of the AV node as well as in thickened fibrous subendocardial regions (results not shown). Evidence of downstream signaling by HIF-1α was demonstrated by the intracellular and extracellular expression of PAI-1 (Figure 1B) (16). Furthermore, intracellular staining of mTOR, a signaling molecule associated with increased collagen synthesis (17), was seen in close proximity to the myofibroblasts (Figure 1C).

thumbnail image

Figure 1. Immunohistologic evidence of hypoxia-induced injury and a profibrosing phenotype in cardiac tissue sections from 2 fetuses with congenital heart block (CHB) and an age-matched healthy control fetus. Shown are images of a longitudinal section through the septum of a 20-week-old CHB-affected heart (A–C), a 22-week-old CHB-affected heart (D–F), and a 23-week-old healthy heart (G–I). Tissue was stained with murine monoclonal antibodies against hypoxia-inducible factor 1α (HIF-1α) (A,D, and G), plasminogen activator inhibitor type 1 (PAI-1) (B, E, and H), and rabbit anti-human mammalian target of rapamycin (mTOR) (C, F, and I); arrows in A and D indicate positive staining (positive results for HIF-1α in D are shown in brown stain, not red). Stains were visualized using anti-mouse IgG alkaline phosphatase (red) or anti-mouse or anti-rabbit specific IgG peroxidase (brown). Cells expressing HIF-1α (evidence of hypoxia) and those expressing PAI-1 and mTOR (profibrotic cells) were detected in the conduction tissue of the hearts from the CHB-affected fetuses, but not the healthy control fetus.

Download figure to PowerPoint

Equivalent findings for the expression of HIF-1α, PAI-1, and mTOR were observed in a 22-week-old CHB-affected heart (Figures 1D–F). In contrast, HIF-1α, PAI-1, and mTOR were not observed in the septum of the normal 23-week-old fetal heart (Figures 1G–I). When an IgG isotype control was used for immunostaining of cardiac tissue from the 20-week-old CHB-affected, 22-week-old CHB-affected, and 23-week-old normal hearts, the results were negative, as expected (results not shown).

Additional supporting evidence of exposure to hypoxic conditions was sought by measurement of erythropoietin in the cord blood from 67 neonates exposed to maternal anti-SSA/Ro antibodies (31 CHB-affected, 36 unaffected). As shown in Figure 2, the median levels of erythropoietin were significantly higher in the CHB-affected neonates than in the unaffected neonates (33.3 mU/ml, interquartile range 68.4 mU/ml versus 20.9 mU/ml, interquartile range 29.3 mU/ml, respectively; P = 0.0085). The 90th percentile of the erythropoietin levels in the antibody-exposed unaffected group was 48.7 mU/ml; this 90th percentile value was exceeded in 3 (8.3%) of 36 blood samples from the unaffected group, as compared with 11 (35.5%) of 31 from the CHB-affected group (P = 0.01).

thumbnail image

Figure 2. Levels of erythropoietin (EPO) in the cord blood of neonates exposed to maternal anti-SSA/Ro antibodies. Evaluation of EPO in cord blood was performed using a specific enzyme-linked immunosorbent assay. Each point represents an individual subject. Median levels of EPO were 33.3 mU/ml (interquartile range 68.4 mU/ml) in the congenital heart block (CHB)–affected neonates compared with 20.9 mU/ml (interquartile range 29.3 mU/ml) in the unaffected healthy control neonates (P = 0.0085, by Mann-Whitney U test). The solid line represents the 90th percentile (median 48.7 mU/ml) of the values in the healthy control group.

Download figure to PowerPoint

Role of hypoxia as an amplification step in fibroblast transdifferentiation.

Having demonstrated in vivo surrogates of hypoxic exposure, in vitro experiments were designed to address functional consequences with regard to amplification of fibrosis. Specifically, we evaluated the effect of hypoxia on the phenotype and gene expression in fibroblasts derived from 2 different organs. Fibroblasts were isolated and cultured in parallel from the heart and the lungs of 3 electively terminated fetuses; both organs were obtained from each of the 3 donors to ensure that any differences were not attributable to individual variability. The fibroblasts were plated in a hypoxia chamber under conditions of <0.01% O2 for 24 hours. The results indicated that LDH was not released into the supernatant, thus confirming that, under these conditions, cells do not undergo necrosis.

Both immunofluorescence staining and immunoblotting revealed that the cardiac fibroblasts expressed SMA in response to hypoxia (Figures 3A versus 3B, and Figure 3D). In contrast, hypoxia did not induce SMA in the pulmonary fibroblasts (each culture having been derived and isolated from the same aborted fetuses as the cardiac fibroblasts) (Figures 3F versus 3G, and Figure 3I). For immunofluorescence and immunoblot analysis, substitution of specific antibodies for a murine IgG isotype control yielded negative results, as anticipated (results not shown).

thumbnail image

Figure 3. Relationship between growth conditions and expression of smooth muscle actin (SMA) in fibroblasts isolated from a fetal heart (A–E) and fetal lung (F–J). Human fetal cardiac fibroblasts (A–E) and human fetal pulmonary fibroblasts (F–J) were prepared as monolayers. Cells were incubated in the absence of treatment (normoxia), or in an atmosphere of <0.01% O2 (hypoxia) for 24 hours, or with transforming growth factor β (TGFβ) (10 ng/ml). For immunofluorescence analysis of SMA, cells were double-stained with Hoechst 33258 and Cy3-labeled anti-SMA (A–C and F–H). For immunoblotting analysis, total protein was prepared by lysis. The lysate was separated on a sodium dodecyl sulfate gel and transferred to nitrocellulose. Specific antigen (SMA [D and I] or tubulin [E and J]) was determined by chemiluminescence assay. In the representative data, the cardiac fibroblasts and lung fibroblasts were obtained from the same donor. The experiment shown is representative of results obtained from 3 different sets of fibroblasts.

Download figure to PowerPoint

Consistent with previous findings (4), the cardiac fibroblasts expressed SMA following exposure to TGFβ (10 ng/ml for 6 hours at 37°C) (Figure 3C), an effect that was not seen in the fibroblasts isolated from the lungs (Figure 3H). In addition to the 3 sets of fibroblasts evaluated, cardiac fibroblasts from 11 other donors showed similar reactivities (results not shown).

Having established that hypoxia promoted the transdifferentiation of the cardiac fibroblasts (as demonstrated by increased expression of SMA), gene expression profiling by Affymetrix arrays was carried out to identify additional genes that might provide clues to the molecular basis for this amplification of the fibroblast contribution to CHB. The expression of 14,500 gene products (a total of 18,400 transcripts) in response to hypoxia was determined. As a baseline value, the data were normalized to the values under normoxia. The list of genes that met the requirements of statistical filtering (as described in Patients and Methods) are available upon request from the corresponding author.

Two of the most highly up-regulated genes included HIG2, an expected footprint of hypoxia that supports growth under low oxygen conditions (18), and adrenomedullin (Figure 4). The latter finding was of particular interest and unexpected, because adrenomedullin is known to raise the levels of cAMP and has been reported to serve as an inhibitor of matrix synthesis (19–22).

thumbnail image

Figure 4. Microarray and quantitative polymerase chain reaction (QPCR) profiling of cardiac fibroblasts (CF) in response to hypoxia. Results are the mean and SEM relative levels of mRNA for vascular endothelial growth factor (VEGF), adrenomedullin (ADM), and hypoxia-inducible protein 2 (HIG2) detected by arrays or quantitative PCR in cardiac fibroblasts from representative fetal tissue after 4 hours of exposure under hypoxic (H) conditions, normalized to time-matched normoxic (N) levels.

Download figure to PowerPoint

Results of the detailed analysis of gene ontology associations for the differentially regulated genes (not shown) further supported the conclusion that there was significant activation of cAMP-specific pathways in response to hypoxia. Other significant functional categories detected included proteins that reflect metabolic adaptation in the maintenance of oxygen homeostasis, such as those involved in glucose transport (e.g., glycolytic enzymes) and angiogenesis.

To confirm the results obtained with microarray, cardiac fibroblasts, in a parallel set of hypoxic conditions, were harvested and directly processed for the preparation of mRNA and examination of cDNA using quantitative PCR. The expression level of the B2M gene was used to normalize for differences in input cDNA. Quantitative PCR experiments confirmed the DNA microarray results (Figure 4). Specifically, levels of transcripts for cardiac fibroblasts (VEGF, adrenomedullin, and HIG2) were markedly increased following exposure to hypoxia.

Modulation of cAMP and regulation of matrix synthesis.

The expression of adrenomedullin, as well as candidates of cAMP-specific pathways induced by hypoxia-treated cardiac fibroblasts, raised the possibility that cAMP might be a key regulating factor. The next set of experiments was designed to evaluate the direct effects of cAMP modulation in the in vitro culture systems. SMA was the initial variable evaluated under the different culture conditions.

Cardiac fibroblasts incubated with the adenylate cyclase inhibitor MDL12,330A exhibited increased expression of SMA, as demonstrated by both immunofluorescence and immunoblotting (Figures 5C and G [lane 3]). Incubation with the cAMP activator forskolin revealed no change in expression of SMA as compared with the baseline culture conditions (Figures 5B and G [lane 2]). In addition, as expected, the cardiac fibroblasts expressed SMA following exposure to TGFβ (10 ng/ml for 6 hours at 37°C) (Figures 5D and G [lane 4]), which was markedly attenuated in the presence of forskolin (Figures 5E and G [lane 5]). Treatment of cardiac fibroblasts with the combination of TGFβ and MDL12,330A did not augment SMA expression beyond the levels that were found with either agonist alone (Figures 5F and G [lane 6]).

thumbnail image

Figure 5. Differential effect of cAMP modulation on expression of smooth muscle actin (SMA) in human fetal cardiac fibroblasts. Cells were prepared in monolayers and incubated in the absence of treatment (control [CNTRL]) (A and G [lane 1]) or in the presence of forskolin (FORSK) (B and G [lane 2]) or MDL12,330A (C and G [lane 3]). In addition, cells were incubated in the presence of transforming growth factor β1 (TGFβ1) (10 ng/ml) (D and G [lane 4]) or TGFβ1 plus cotreatment with forskolin (E and G [lane 5]) or cotreatment with MDL12,330A (F and G [lane 6]). Cells were double-stained with Hoechst 33258 and primary antibodies (monoclonal antibodies to SMA) followed by the species-specific secondary anti-IgG tetramethylrhodamine isothiocyanate (A–F). In addition, for each condition, SMA levels were determined by immunoblot analysis (G). Results are representative of those obtained from 11 different donors of cardiac fibroblasts.

Download figure to PowerPoint

thumbnail image

Figure 6. Differential effect of cAMP modulation on expression of mammalian target of rapamycin (mTOR), plasminogen activator inhibitor type 1 (PAI-1), and type I collagen (COL 1) in human fetal fibroblasts isolated from the heart and lung. Human fetal cardiac fibroblasts and human fetal pulmonary fibroblasts were prepared in monolayers and incubated in the absence (control [CNTRL]) or presence of MDL12,330A (30 μM). Cells were double-stained with Hoechst 33258 and monoclonal antibodies to mTOR (A), PAI-1 (B), or type I collagen (C), followed by the species-specific secondary anti-IgG tetramethylrhodamine isothiocyanate. Immunofluorescence results are representative of those obtained from 3 different sets of fibroblasts. In addition, for each condition, mTOR and PAI-1 levels were determined by immunoblot analysis, as described in Figure 3D.

Download figure to PowerPoint

Further analyses of a profibrosing phenotype comprised assessment of the expression of mTOR, PAI-1, and type I collagen in cardiac fibroblasts. In parallel with SMA expression, cardiac fibroblasts incubated with MDL12,330A exhibited increased expression of mTOR, PAI-1, and type I collagen compared with that in the baseline culture conditions, as assessed by immunofluorescence and immunoblotting (Figures 6A–C [lane 1 versus lane 2] and 6D). In contrast, treatment of lung fibroblasts with MDL12,330A did not increase expression of mTOR, PAI-1, or type I collagen (Figures 6A–C [lane 3 versus lane 4]).

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

To date, a reasonable explanation for the rarity of CHB remains a challenging research issue. While substantial evidence exists that anti-SSA/Ro antibodies are a necessary factor, the discordance of disease in identical twins is an observation difficult to reconcile. Given the consideration that a stressful in utero environment might differentially affect one fetus versus another, hypoxia was addressed as an accelerating factor that might tip the balance from vulnerability to full expression of disease.

Based on the intriguing observations of rapid scarring of the AV node, the concomitant or even isolated findings of endocardial fibroelastosis, and the susceptibility of the heart to CHB compared with other organs, we selected fetal cardiac and lung fibroblasts as the focus of this study. The results presented herein support hypoxia as an environmental stress that potentially affects the distal component of the pathologic cascade. In keeping with our own hypothesis previously reported (23), and with genetic studies on the increased frequency of a profibrosing TGFβ polymorphism in CHB-affected fetuses (24, 25), it is envisioned that hypoxia may amplify the injurious effects of scarring cytokines, such as TGFβ, that are secreted by infiltrating macrophages during the clearance of opsonized anti-SSA/Ro–bound cardiocytes.

It is particularly challenging to provide proof of concept with regard to the hypoxia hypothesis, because in only 20% of fetuses CHB is fatal, and in these cases, death is often weeks after the detection of bradycardia, making it difficult to identify relevant pathologic footprints. The availability of an electively terminated fetal heart for immunohistologic evaluation within days of the echocardiographic diagnosis of CHB (20 weeks) and another fetus dying close to the time of diagnosis (22 weeks) provided us with a rare opportunity to assess very proximate biologic events. Although it is fully acknowledged that a decrease in cardiac output itself could be the cause of hypoxia, in the electively terminated fetus studied herein there were no signs of low cardiac output due to myocardial dysfunction at the detection of CHB.

The accumulation of HIF-1α, a well-characterized transcription factor complex that regulates hypoxia-driven gene expression (26, 27), may represent a causative factor. In addition, expression of mTOR was recently advanced as a mediator of fibrosis in scleroderma (17). Notably, the HIF-1α and mTOR were localized to regions in which we have previously observed TGFβ staining and nuclear translocation of SMAD2 (3). Expression of HIF-1α in the second CHB case is more difficult to interpret, since this fetus had a severe cardiomyopathy that resulted in death within 2 weeks of the diagnosis of CHB; therefore, attribution of hypoxia to the initial injury is more ambiguous. HIF-1α was not detected in the electively terminated normal heart. Testing the amplifying effect of hypoxia may be of interest in an animal model, since attempts to generate anti-SSA/Ro–mediated heart block in pups has yielded observations of predominantly first-degree block (28) and a low penetrance of third-degree block (28–30).

A hypoxic stimulus results in an increase in erythropoietin levels, to enhance red cell oxygen-carrying capacity. Cord levels of this glycoprotein, which does not cross the placenta, are an indicator of tissue oxygenation and reflect production solely by the fetus. In contrast to umbilical arterial pH changes, which reflect acute hypoxia, elevations of erythropoietin are considered to indicate prolonged fetal hypoxia. Relevant to the timing of CHB, increased fetal production of erythropoietin has been reported to occur from 20 weeks' gestation (31). Thus, elevated cord blood levels could represent ongoing hypoxic conditions that were initiated months earlier and have been sustained.

In consideration of erythropoietin as a biomarker of hypoxia, it is notable that in several studies, a negative association with birth weight has been observed (32). Pertinent to the consideration of hypoxia as an amplifying factor in CHB (as well as the discordance of disease in twins exposed to identical maternal antibodies) is the report that in 5 sets of twins (not related to neonatal lupus) in whom birth weights differed by more than 25%, the mean erythropoietin concentration was significantly higher in the smaller twin than in the larger twin (32). Arguably, in the situation of CHB, it is equally plausible that hypoxia was not causal but, rather, a consequence of the disease once established. However, based on the data presented herein, CHB per se does not invariably increase the cord blood level of erythropoietin. This makes sense, since hypoxia is hypothesized to be an amplifying, but not requisite, factor in the pathogenesis of CHB. Given a sufficiently vigorous inflammatory component, scarring would be predicted to proceed in normoxic conditions. In the absence of fetal sampling during gestation (impractical for research purposes), the finding of elevated cord blood levels of erythropoietin conceptually supports the in vitro findings presented herein, but should be interpreted with consideration of potential confounders.

Adrenomedullin, a 52–amino-acid vasodilator protein that up-regulates cAMP in target cells including fibroblasts (21, 33), was identified via the array approach. Adrenomedullin has previously been reported to be synthesized by endothelial cells, healthy renal cells, and renal carcinoma cells exposed to hypoxia (34–36). The increased transcription of adrenomedullin following exposure to hypoxia provided the link to cAMP, but generated a potential paradox in that the net effect of hypoxia was to augment fibrosis. Nguyen and Claycomb demonstrated that in adult rat ventricular cardiac myocytes, hypoxia induced transcription of the adrenomedullin genes by HIF-1α (37). Findings suggestive of the protective effects of endogenous adrenomedullin on cardiac fibrosis were obtained in mice heterozygous for an adrenomedullin-null mutation, since left ventricular wall thickness and perivascular fibrosis following aortic constriction were worse in these mice compared with wild-type mice (38). One plausible explanation for the results reported herein is that in the human fetal cardiac fibroblast, adrenomedullin is expressed in an attempt to protect against fibrosis, but is ineffective in repletion of sufficient cAMP.

With regard to the effects of both hypoxia and TGFβ, changes in cAMP levels appear to be highly influential in the regulation of cardiac fibroblast transdifferentiation. Transformation of cardiac fibroblasts to myofibroblasts, characterized by expression of SMA and the production of extracellular matrix components, is a pivotal event in the remodeling of connective tissue. Exposure of human fetal cardiac fibroblasts to hypoxia resulted in the expression of a scarring phenotype, equivalent to the effect observed after coincubation with TGFβ. Raising the level of cAMP inhibited this effect. In support of these findings, Swaney et al recently reported that increased cAMP levels modulate the transformation of adult rat cardiac fibroblasts to myofibroblasts; specifically, forskolin inhibited TGFβ-induced SMA expression and collagen synthesis (21).

One theory for the resistance to a scarring phenotype that has been observed in the pulmonary fibroblast is that this organ is in a state of chronic hypoxia, given that blood flow to the lungs is not important until birth. These findings suggest that pleiotropic responses are cell-type–specific, as has been previously reported (39). Perhaps differential susceptibility to changes in cAMP accounts, in part, for the predominance of the heart as the target organ injured by maternal anti-SSA/Ro antibodies. Thus, in addition to an inflammatory injury, the inherent inability of cardiac tissue to maintain oxygen homeostasis when challenged may tip the balance toward scar formation.

It is readily acknowledged that translation of our in vitro findings to human pregnancy and CHB is limited by the available data. However, hypoxia does occur during pregnancy and can originate either in the mother or at the level of the placenta, since oxygen exchange between the mother and fetus occurs via the placenta. Although the generally accepted causes of maternal hypoxia, such as hypertension, diabetes, infection, anemia, and smoking, have not been systematically evaluated during pregnancies in which the anti-SSA/Ro–positive mother has a fetus with a diagnosis of CHB, changes in oxygen during the period of vulnerability (at 16–24 weeks' gestation) appear to be plausible.

Our data thus support the novel hypothesis that hypoxia is a potential environmental stress factor that is capable of amplifying the myofibroblast transdifferentiation of a fetal cardiac fibroblast that has already been subjected to an inflammatory injury initiated by circulating maternal anti-SSA/Ro antibodies. In the process of normal repair, the fibroblast is expected to mediate a role involving matrix synthesis that is associated with reversible wounding without scar formation. In this scenario, hypoxia causes an increase in HIF-1α, which induces the expression of SMA, but at the same time increases cAMP. Thus, hypoxia under physiologic circumstances promotes a matrix-promoting process and a molecule that restrains this process. In a pathologic scenario, the environment (e.g., hypoxia) promotes exuberant matrix synthesis that is irreversible (a fibrotic process). This may occur in the CHB-affected fibroblast, which is primed by antecedent inflammatory events and already expressing HIF-1α, as evidenced by histologic findings. Increasing focus on the causes and prevention of abnormal oxygen homeostasis may provide a new direction for research on the pathogenesis of CHB.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Dr. Clancy 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.

Study design. Clancy, Buyon.

Acquisition of data. Zheng, O'Mahony, Izmirly, Gardner.

Analysis and interpretation of data. Clancy, Buyon.

Manuscript preparation. Clancy, Buyon.

Statistical analysis. Zavadil.

Acknowledgements

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

We thank the New York University Cancer Institute genomics staff for performing the DNA array analysis, Dr. Brad Poulos at the Human Fetal Tissue Repository of Albert Einstein College of Medicine (Bronx, NY) for providing the normal tissue samples, and Dr. Mimi Kim at Albert Einstein College of Medicine for providing statistical advice.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  • 1
    Buyon JP, Clancy RM. Neonatal lupus. In: WallaceDJ, HahnBH, editors. Dubois' Lupus Erythematosus. 7th ed. Philadelphia: Lippincott Williams & Wilkins; 2006. p. 105880.
  • 2
    Brucato A, Frassi M, Franceschini F, Cimaz R, Faden D, Pisoni MP, et al. Risk of congenital complete heart block in newborns of mothers with anti-Ro/SSA antibodies detected by counterimmunoelectrophoresis: a prospective study of 100 women. Arthritis Rheum 2001; 44: 18325.
  • 3
    Clancy RM, Kapur RP, Molad Y, Askanase AD, Buyon JP. Immunohistologic evidence supports apoptosis, IgG deposition, and novel macrophage/fibroblast crosstalk in the pathologic cascade leading to congenital heart block. Arthritis Rheum 2004; 50: 17382.
  • 4
    Clancy RM, Askanase AD, Kapur RP, Chiopelas E, Azar N, Miranda-Carus ME, et al. Transdifferentiation of cardiac fibroblasts: a fetal factor in anti-SSA/Ro-SSB/La antibody-mediated congenital heart block. J Immunol 2002; 169: 215663.
  • 5
    Miranda-Carus ME, Askanase AD, Clancy RM, Di Donato F, Chou TM, Libera MR, et al. Anti-SSA/Ro and anti-SSB/La autoantibodies bind the surface of apoptotic fetal cardiocytes and promote secretion of TNF-α by macrophages. J Immunol 2000; 165: 534551.
  • 6
    Askanase AD, Friedman DM, Copel J, Dische MR, Dubin A, Starc TJ, et al. Spectrum and progression of conduction abnormalities in infants born to mothers with anti-SSA/Ro-SSB/La antibodies. Lupus 2002; 11: 14551.
  • 7
    Millen J, MacLean MR, Houslay MD. Hypoxia-induced remodelling of PDE4 isoform expression and cAMP handling in human pulmonary artery smooth muscle cells. Eur J Cell Biol 2006; 85: 67991.
  • 8
    Levy O, Coughlin M, Cronstein BN, Roy RM, Desai A, Wessels MR. The adenosine system selectively inhibits TLR-mediated TNF-α production in the human newborn. J Immunol 2006; 177: 195666.
  • 9
    Tseng CE, Miranda E, Di Donato F, Boutjdir M, Rashbaum W, Chan EK, et al. mRNA and protein expression of SSA/Ro and SSB/La in human fetal cardiac myocytes cultured using a novel application of the Langendorff procedure. Pediatr Res 1999; 45: 2609.
  • 10
    Gardner LB, Li F, Yang X, Dang CV. Anoxic fibroblasts activate a replication checkpoint that is bypassed by E1a. Mol Cell Biol 3003; 23: 903245.
  • 11
    Irizarry RA, Bolstad BM, Collin F, Cope LM, Hobbs B, Speed TP. Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res 2003; 31: e15.
  • 12
    Tusher VG, Tibshirani R, Chu G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci U S A 2001; 98: 511621.
  • 13
    Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2math image method. Methods 2001; 25: 4028.
  • 14
    Buyon JP, Hiebert R, Copel J, Craft J, Friedman D, Katholi M, et al. Autoimmune-associated congenital heart block: demographics, mortality, morbidity and recurrence rates obtained from a national neonatal lupus registry. J Am Coll Cardiol 1998; 31: 165866.
  • 15
    Friedman DM, Kim MY, Copel A, Davis C, Buyon JP. Utility of cardiac monitoring in fetuses at risk of congenital heart block: the PR interval and dexamethasone evaluation (PRIDE) prospective study [abstract]. Arthritis Rheum 2006; 54 Suppl 9: S689.
  • 16
    Li X, Kimura H, Hirota K, Kasuno K, Torii K, Okada T, et al. Synergistic effect of hypoxia and TNF-α on production of PAI-1 in human proximal renal tubular cells. Kidney Int 2005; 68: 56983.
  • 17
    Shegogue D, Trojanowska M. Mammalian target of rapamycin positively regulates collagen type I production via a phosphatidylinositol 3-kinase-independent pathway. J Biol Chem 2004; 279: 2316675.
  • 18
    Togashi A, Katagiri T, Ashida S, Fujioka T, Maruyama O, Wakumoto Y, et al. Hypoxia-inducible protein 2 (HIG2), a novel diagnostic marker for renal cell carcinoma and potential target for molecular therapy. Cancer Res 2005; 65: 481726.
  • 19
    Duncan MR, Frazier KS, Abramson S, Williams S, Klapper H, Huang X, et al. Connective tissue growth factor mediates transforming growth factor β-induced collagen synthesis: down-regulation by cAMP. FASEB J 1999; 13: 177486.
  • 20
    Saltzman LE, Moss J, Berg RA, Hom B, Crystal RG. Modulation of collagen production by fibroblasts: effects of chronic exposure to agonists that increase intracellular cyclic AMP. Biochem J 1982; 204: 2530.
  • 21
    Swaney JS, Roth DM, Olson ER, Naugle JE, Meszaros JG, Insel PA. Inhibition of cardiac myofibroblast formation and collagen synthesis by activation and overexpression of adenylyl cyclase. Proc Natl Acad Sci U S A 2005; 102: 43742.
  • 22
    Thalacker FW, Nilsen-Hamilton M. Opposite and independent actions of cyclic AMP and transforming growth factor β in the regulation of type 1 plasminogen activator inhibitor expression. Biochem J 1992; 287(Pt 3): 85562.
  • 23
    Buyon JP, Clancy RM. Maternal autoantibodies and congenital heart block: mediators, markers, and therapeutic approach. Semin Arthritis Rheum 2003; 33: 14054.
  • 24
    Clancy RM, Backer CB, Yin X, Kapur RP, Molad Y, Buyon JP. Cytokine polymorphisms and histologic expression in autopsy studies: contribution of TNF-α and TGF-β 1 to the pathogenesis of autoimmune-associated congenital heart block. J Immunol 2003; 171: 325361.
  • 25
    Cimaz R, Borghi MO, Gerosa M, Biggioggero M, Raschi E, Meroni PL. Transforming growth factor β1 in the pathogenesis of autoimmune congenital complete heart block: lesson from twins and triplets discordant for the disease. Arthritis Rheum 2006; 54: 3569.
  • 26
    Sanchez-Elsner T, Botella LM, Velasco B, Corbi A, Attisano L, Bernabeu C. Synergistic cooperation between hypoxia and transforming growth factor-β pathways on human vascular endothelial growth factor gene expression. J Biol Chem 2001; 276: 3852735.
  • 27
    Semenza GL. HIF-1 and human disease: one highly involved factor. Genes Dev 2000; 14: 198391.
  • 28
    Salomonsson S, Sonesson SE, Ottosson L, Muhallab S, Olsson T, Sunnerhagen M, et al. Ro/SSA autoantibodies directly bind cardiomyocytes, disturb calcium homeostasis, and mediate congenital heart block. J Exp Med 2005; 201: 117.
  • 29
    Mazel JA, El-Sherif N, Buyon J, Boutjdir M. Electrocardiographic abnormalities in a murine model injected with IgG from mothers of children with congenital heart block. Circulation 1999; 99: 19148.
  • 30
    Miranda-Carus ME, Boutjdir M, Tseng CE, DiDonato F, Chan EK, Buyon JP. Induction of antibodies reactive with SSA/Ro-SSB/La and development of congenital heart block in a murine model. J Immunol 1998; 161: 588692.
  • 31
    Thilaganathan B, Salvesen DR, Abbas A, Ireland RM, Nicolaides KH. Fetal plasma erythropoietin concentration in red blood cell-isoimmunized pregnancies. Am J Obstet Gynecol 1992; 167: 12927.
  • 32
    Maier RF, Obladen M, Kattner E, Natzschka J, Messer J, Regazzoni BM, et al, the European Multicenter rhEPO Study Group. High-versus low-dose erythropoietin in extremely low birth weight infants. J Pediatr 1998; 132: 86670.
  • 33
    Oie E, Vinge LE, Yndestad A, Sandberg C, Grogaard HK, Attramadal H. Induction of a myocardial adrenomedullin signaling system during ischemic heart failure in rats. Circulation 2000; 101: 41522.
  • 34
    Scheurer SB, Rybak JN, Rosli C, Neri D, Elia G. Modulation of gene expression by hypoxia in human umbilical cord vein endothelial cells: a transcriptomic and proteomic study. Proteomics 2004; 4: 173760.
  • 35
    Fujita Y, Mimata H, Nasu N, Nomura T, Nomura Y, Nakagawa M. Involvement of adrenomedullin induced by hypoxia in angiogenesis in human renal cell carcinoma. Int J Urol 2002; 9: 28595.
  • 36
    Sandner P, Hofbauer KH, Tinel H, Kurtz A, Thiesson HC, Ottosen PD, et al. Expression of adrenomedullin in hypoxic and ischemic rat kidneys and human kidneys with arterial stenosis. Am J Physiol Regul Integr Comp Physiol 2004; 286: R94251.
  • 37
    Nguyen SV, Claycomb WC. Hypoxia regulates the expression of the adrenomedullin and HIF-1 genes in cultured HL-1 cardiomyocytes. Biochem Biophys Res Commun 1999; 265: 3826.
  • 38
    Niu P, Shindo T, Iwata H, Iimuro S, Takeda N, Zhang Y, et al. Protective effects of endogenous adrenomedullin on cardiac hypertrophy, fibrosis, and renal damage. Circulation 2004; 109: 178994.
  • 39
    Denko NC, Fontana LA, Hudson KM, Sutphin PD, Raychaudhuri S, Altman R, et al. Investigating hypoxic tumor physiology through gene expression patterns. Oncogene 2003; 22: 590714.