More than just a small left ventricle: the right ventricular fibroblast and ECM in health and disease

Abstract

Fibroblasts intricately organize and regulate the extracellular matrix (ECM) in cardiac health and disease. Excess deposition of ECM proteins causes fibrosis, resulting in disrupted signaling conduction and contributing to the development of arrhythmias and impaired cardiac function. Fibrosis is causally involved in cardiac failure in the left ventricle (LV). Fibrosis likely occurs in right ventricle (RV) failure, yet mechanisms remain unclear. Indeed, RV fibrosis is poorly understood with mechanisms often extrapolated from the LV to the RV. However, emerging data suggest that the LV and RV are distinct cardiac chambers and differ in regulation of the ECM and response to fibrotic stimuli. In the present review, we will discuss differences in ECM regulation in the healthy RV and LV. We will discuss the importance of fibrosis in the development of RV disease in pressure overload, inflammation, and aging. During this discussion, we will highlight mechanisms of fibrosis with respect to the synthesis of ECM proteins while acknowledging the importance of considering collagen breakdown. We will also discuss current knowledge of antifibrotic therapies in the RV and the need for additional research to help delineate the shared and distinct mechanisms of RV and LV fibrosis.

AN INTRODUCTION TO THE RIGHT VENTRICLE

The right (RV) and left (LV) ventricles possess distinct anatomical and physiological characteristics that allow them to pump to the lungs and periphery, respectively. The triangular thinner-walled RV works against low pressure in the pulmonary circuit to deliver blood to the lungs for oxygenation. Reciprocally, the LV must work against higher pressures to pump throughout the peripheral systemic vasculature. Failure of either or both ventricles results in heart failure (HF), a burdensome clinical condition with poor patient outcomes. One in three individuals over the age of 55 will develop HF and survival after diagnosis remains poor. Either ventricle can fail, but LV failure is the most common. However, because the circulatory system is a closed loop, LV dysfunction often impacts the RV. Indeed, RV failure is mostly caused by the backward transmission of high LV filling pressures to the RV, consequently causing pulmonary hypertension (PH). PH causes RV failure, collectively referred to as cor pulmonale (“pulmonary heart”). Cor pulmonale by definition develops independent of the LV such as from idiopathic pulmonary arterial hypertension, hypoxic lung disease-induced PH, and chronic thromboembolic PH (1). Right-sided heart failure may also occur independent of PH because of myocarditis and ischemia (2). Mechanisms of HF that have been the target of decades of clinical and preclinical research have been gleaned predominantly from the LV. However, RV function predicts mortality in age-related HF (3) and PH (4), highlighting the importance of understanding the underlying biological, physiological, and molecular characteristics of the RV. Throughout the past few decades of cardiac research, the RV has been referred to as the “forgotten ventricle” (5), the “dark side of the moon” (6), or a “mere bystander” with respect to its contributions to cardiac function. With this bias in mind, it is not surprising that limited relative efforts have focused on the RV compared with the LV. The assumption that the RV is a smaller version of the LV has led to a paucity of data on the RV and has precluded therapeutic advancements for RV disease. The significance of the RV in biomedicine and the need for RV studies is exemplified by a recent statement from a National Heart, Lung, and Blood Institute working group stating “RV failure cannot be understood simply by extrapolating data and experience from LV failure. The RV is different from the LV” (7). For the focus of this review, we will discuss what is known about how the ventricles differ with respect to fibrosis and how fibrotic signaling may be dysregulated with age and disease in the RV.

THE ECM IN THE HEALTHY DEVELOPING RV AND LV

The myocardium is composed of a heterogeneous mix of cardiomyocytes, fibroblasts, and other cell types functioning in a cohesive manner to maintain homeostasis. Cardiac fibroblasts represent ∼13% of all cells by number in the healthy heart and are the predominant cell type responsible for the regulation of the extracellular matrix (ECM) (8). The ECM is a dynamic three-dimensional network that acts as a signaling hub and provides mechanical support to help maintain sarcomere length and collagen-myocyte tethering (9). The primary ECM protein produced by fibroblasts that is also most associated with fibrosis is collagen. In the myocardium, collagen types I and III are the most abundant, followed by less abundant types IV, V, and VI (10). Collagens I and III are fibrillary and have a clear structural role. Collagen type 1 (Col1) is responsible for building strong, thick fibers and accounts for ∼85% of total myocardial collagens. The thinner, more elastic fibril collagen type III (Col3) accounts for ∼11% of total collagen content. Fibril collagens and other ECM proteins are regulated by matrix metalloproteinases (MMPs), a family of proteases responsible for the breakdown of ECM proteins (11). Diverse in their substrates and actions, MMPs degrade collagens and other ECM proteins. The proteolytic function of MMPs is tightly controlled by their inhibitors tissue inhibitors of metalloproteinases (TIMPs) that inhibits MMPs with varying specificity and affinities (12). A tightly controlled balance between ECM protein synthesis as well as MMP and TIMP activity is essential to sustain the integrity of the ECM to preserve a healthy matrix.

The RV and LV derive from different heart fields, with the RV developing from the secondary heart field and the LV from the primary heart field (13). Perhaps not unexpectedly given these distinct embryological origins, fibroblasts in the two ventricles also derive from distinct locations, with LV fibroblasts from the epicardium and endocardium, whereas RV fibroblasts only derive from the epicardium (14). Recent single-cell sequencing data suggest that while myocytes are the most different cell type between the RV and the LV, fibroblasts also differ between the two ventricles. Enrichment analysis in fibroblast cell populations demonstrated 25 left- and 51 right-sided specific genes (15). Gene expression of ECM regulators differs between the ventricles (16). Differences in collagen content between the healthy RV and LV have been debated for the past several decades. Early reports comparing development of the RV and LV suggested that while collagen content was the same in the RV and LV before birth, in the first few months of development, RV collagen content increased at a greater rate than the LV, resulting in higher collagen in the mature RV (Fig. 1). The authors posited that the greater magnitude of hypertrophy in the LV postbirth (four times that of the RV) was responsible for the smaller percent collagen compared with the RV (17, 18). This finding has been repeated in some models such as the healthy rat heart that has ∼20% more collagen in the RV either by hydroxyproline or percent collagen (19, 20). However, in other models, such as young mice, RV and LV fibrotic percentages are similar (21, 22). Although some of the discrepancy may be due to species or methodological differences, it is also possible that anatomic differences of measurement explain if the LV and RV differ with respect to baseline collagen content. Although collagen concentration was reported to be higher in the RV compared with the LV + septum, RV collagen content was not higher than the LV free wall in young healthy male rats (23). It is not fully clear yet how the RV and LV differ with respect to cardiac ECM composition. Although initially similar at birth, fibrotic and hypertrophic mechanisms during development likely separate the RV and LV. Understanding these mechanisms with emerging high throughput and other innovative technologies will hopefully begin to elucidate these important differences.

Figure 1.

Proposed differences in the healthy extracellular matrix (ECM) in the left ventricle (LV) and right ventricle (RV) due to developmental changes in chamber composition. At birth, cellular composition of the RV and LV is likely similar. However, during development, more rapid growth of myocytes in the LV results in a lower composition of fibroblasts and collagen compared with the RV. The RV also has higher proportion of immune cells than the LV that may contribute to different fibrosis signaling pathways and collagen deposition.

FIBROSIS IN THE DISEASED RV AND LV

When exposed to injury, the heart releases growth factors, inflammatory cytokines, and other signaling proteins that result in the activation of quiescent fibroblasts into myofibroblasts. Myofibroblasts maintain a procollagen deposition phenotype with migratory and secretory properties, leading to the accumulation of ECM proteins (24). Initially intended to be reparative to help maintain cardiac structure and function as described earlier, in response to chronic stress, excess ECM is deposited, resulting in the development of fibrosis. Net accrual of ECM proteins could be due to increased ECM protein deposition, decreased ECM degradation, or a combination thereof. To date, the bulk of fibrotic mechanisms and what will be presented in this review have aimed at understanding mechanisms of ECM protein synthesis. However, there is growing acknowledgment that removal of ECM proteins is equally important, particularly with the discovery of posttranslational modifications of ECM proteins. Regardless of mechanism, excess ECM proteins create a stiffer, less conductive myocardium, contributing to both systolic and diastolic dysfunction and disease progression (25, 26). Fibrotic protein patterns are dependent upon disease. Regardless of where the excess ECM is deposited, however, myofibroblasts are responsible. Myofibroblasts can also synthesize periostin (POSTN) and other nonstructural proteins that bind to and impact the ECM network by regulating and transducing signal cascades as well as the activity of proteases, cytokines, and other growth factors (27). Myofibroblasts express contractile proteins such as α-smooth muscle actin (α-SMA), which upregulates overall contractile activity (28), contributing to a stiffer, less relaxed myocardium. Thus, myofibroblasts are responsible for numerous complex signaling pathways reinforcing myofibroblast proliferation and activation making them a hallmark of fibrosis.

In addition to a higher quantity of collagen, the phenotype of the collagen proteins is also altered during disease. Posttranslational modifications impact collagen mechanical properties. Lysyl oxidase and advanced glycation end products cross-link collagen (29). When collagen is cross-linked, its tensile strength increases and it becomes resistant to degradation (30). Myocardial collagen cross-linking associates with diastolic dysfunction in heart failure with preserved ejection fraction (31, 32) and with adverse outcomes in experimental models of hypertension (33), aortic banding (34), and LV volume overload (35). Collagen cross-linking but not collagen content was associated with elevated LV filling pressures in patients with hypertensive HF (36), suggesting that the degree of collagen cross-linking determines LV stiffness more than the number of collagen fibers.

We note for the purposes of this introduction on cellular mechanisms of fibrosis that the majority of these papers were drawn from LV literature. While we presume that RV myofibroblasts act similarly to deposit collagen and its cross-linked form, the specific mechanisms are still unclear and will be the focus of additional discussion.

Methods to Quantify RV and LV Fibrosis

Cardiac fibrosis is quantified in clinical and preclinical models with varying degrees of completeness and accessibility. The most common method to quantify fibrosis is by histology. Several histological protocols including trichrome and Picrosirius red stain collagen in tissue sections. However, the pattern of collagen stain is highly variable between ventricles and one subsection of cardiac tissue likely has low representativeness of the whole ventricle. It is difficult to detect collagen deposition in a small sample when it may be localized in specific portions of the heart (37). Hydroxyproline assays to some extent circumvent this issue by homogenizing a larger chunk of ventricle to quantify the hydroxylation of proline within the collagen molecule (38). Hydroxyproline is highly specific to collagen and thus reflects the amount of collagen in a tissue. Histological or hydroxyproline quantification of collagen is often accompanied by quantification of the expression of profibrotic genes POSTN, α-SMA, and collagens I and III among others. Recent technological advances have aimed to quantify collagen in the heart in vivo and are suited for noninvasive use in human subjects. Late gadolinium enhancement affords visual quantification of myocardial fibrosis by the retention of gadolinium in the myocardium. Gadolinium spreads through the extracellular space but does not enter intact cells. The washout of gadolinium is altered in tissues with high collagen content, thus quantifying fibrotic areas (39). Speckle-tracking echocardiography is emerging as a new indirect detection method for myocardial fibrosis. Using image-processing algorithms, small stable myocardial footprints or “speckles” are identified within an area and are then tracked to determine distances between other speckles or spatiotemporal displacement. Two-dimensional and 3-D speckle tracking echocardiography has been used to correlate histopathological findings of myocardial fibrosis with echocardiographic indices in patients with advanced HF that underwent heart transplant (40). Together, methods to measure cardiac function can be paired with detection methods to help elucidate how fibrosis affects cardiac function in disease.

Because myofibroblasts are responsible for the excess ECM deposition and fibrosis, recent efforts have focused on understanding mechanisms of activation and the phenotype of these cells. These in vitro studies are performed by isolation and culture of cardiac (typically LV) fibroblasts, followed by introduction of a differentiating agent such as transforming growth factor-β (TGF-β) that stimulates quiescent cardiac fibroblasts into myofibroblasts. Subsequent experiments then investigate characteristics of myofibroblasts such as proliferative and contractile capacity. Although it is conceptually feasible to isolate RV fibroblasts, fewer studies have done so. Tian et al. isolated RV fibroblasts from male rats with and without RV dysfunction. The authors found profibrotic epigenetic mechanisms as well as RV fibroblast proliferation and collagen production was higher in RV fibroblasts from diseased compared with control rats (41). RV fibroblasts from pulmonary hypertension-exposed neonatal calves also maintain a transcriptomic signature with gene expression related to ECM remodeling and cytokine expression (42). Genetic models designed to study fibroblast-specific mechanisms of disease have taken advantage of some putative fibroblast-enriched markers such as α-SMA or POSTN (43). However, to date, these genetic models have not been used in RV models of disease, likely in part, because of expression of these genes in non-RV populations (i.e., the pulmonary vasculature). Fibroblast populations are heterogeneous, making it difficult to develop cardiac fibroblast- or myofibroblast-specific markers. Therefore, there are no specific molecular markers for cardiac fibroblasts and myofibroblasts, let alone RV cardiac fibroblasts and myofibroblasts. While mechanistic studies of the RV fibroblast would move the field as would RV fibroblast-specific identifiers, to date these approaches have been very limited. Future approaches such as single-cell RNAseq may identify more sensitive fibroblast and myofibroblast markers, possibly even identifying those that permit distinction between RV and LV fibroblasts, a huge unmet need.

The tools and methods we have discussed thus far have been used to measure collagen quantity or mechanisms of collagen deposition. However, as discussed above, fibrosis is a balance of both synthesis and degradation. To date, most assessments of collagen breakdown have used some combination of protein or gene expression of MMPs or collagen cross-linking mediators such as lysyl oxidase and advanced glycation end products (23, 36, 44). Although these regulators of collagen solubility and degradation are important, they do not directly measure the rates of breakdown. However, newer methods are emerging that directly quantify the rates of collagen protein degradation using isotopes. Deuterium oxide is a stable isotope that permits the quantification of long-term measurements of protein turnover. Recently, deuterium oxide-tracing methods have been applied to quantify collagen protein synthesis and breakdown in skeletal muscle (45). When used to measure ECM turnover in young and aged rats, the authors found that a larger percentage of the aged muscle collagen pool was not turning over, suggesting resistance to degradation. Using a mathematical approach that accounts for the collagen protein pool in young and aged animals, they demonstrated that collagen protein breakdown slows with age, contributing to the elevated collagen content in the aged muscle. To our knowledge, similar approaches have not yet been applied in the heart, much less the RV. Application of these and similar methods to understand mechanisms of collagen degradation will help elucidate additional novel mechanisms of cardiac fibrosis.

Fibrosis in RV Failure

Fibrosis is a primary contributor to LV dysfunction. The severity of LV fibrosis has been linked to mortality and patient outcomes in patients with the lowest myocardial fibrosis showing the best clinical outcomes (46). The association between fibrotic remodeling and ventricular function holds true in arrhythmogenic right ventricular cardiomyopathy. Arrhythmogenic right ventricular cardiomyopathy is characterized by fibro-fatty replacement of RV myocytes, leading to arrhythmia and RV failure (47). However, the link between RV function and fibrosis in other forms of RV-centric disease is less clear (48). Though the RV and LV appear to adapt to stress via some similar mechanisms, further investigation is required to understand ventricle-specific fibrotic remodeling. In the following section, we will discuss fibrosis in the diseased RV, comparing mechanisms of RV fibrosis to the robust literature that exists for LV. Because the interest of our group and the focus of the current review is on the RV, we have not exhaustively discussed LV fibrosis and refer the reader to excellent reviews referenced below. A summary of fibrotic mechanisms and their contributions to RV and LV function in pressure overload, inflammation, and aging is presented in Table 1.

Table 1.

Summary of mechanisms of RV and LV fibrosis in response to pressure overload, inflammation, and aging

	LV	RV	Summary
Pressure overload	Increase in collagen content (26, 49) and I:III ratio (28, 50) Increase in collagen cross-linking (32) associated with LV stiffness and diastolic dysfunction. Direct links between fibrotic remodeling and LV dysfunction (32, 46, 51).	Increase in collagen content and activation of profibrotic gene expression (50, 52–59) that sometimes (25) but not always associate with impaired RV function (26). Robust and consistent activation of POSTN (60, 61). Decreased collagen degradation (56).	Deposition of collagen occurs in pressure overloaded RV and LV. Similar magnitude of pressure overload results in greater activation of fibrotic gene signaling in RV than LV. RV collagen cross-linking requires further investigation as does the association of RV fibrosis with function.
Inflammation	Injury stimulates immune cell migration (62) and myofibroblast differentiation (63). Immune cell loss lowers collagen synthesis (64) and myofibroblast infiltration (65). Increase in cytokine activity promotes collagen deposition (65) while loss of attenuates deposition (66).	Higher macrophages and dendritic cells (16). Immune cells and cytokine activity elevated in PH (67,68). Collagen content not linked to inflammation in IPAH and SScPAH (69).	Immune cell composition differs in the healthy RV compared with the LV. RV and LV upregulate cytokines with disease. Fibrosis might not be as directly associated with inflammation in PH-induced RV fibrosis.
Aging	Increase in collagen content and shift from collagen type III to I (70) with decreased or unchanged profibrotic gene expression (71). Fibrosis correlates with diastolic dysfunction (72,73). Increase in collagen cross-linking (72, 74). Decrease in collagen degradation gene expression (75).	Age-associated collagen deposition (22, 23) with higher expression of profibrotic genes. Negligible collagen cross-linking (23). Unknown how collagen degradation and associated signaling changes with age.	Collagen deposition increases with age in both ventricles, but to a larger extent in the RV (76) and likely via distinct mechanisms. Links between RV fibrosis and function not yet clear, in part due to lack of consensus on changes in RV function with age.

Current understanding of mechanisms of fibrosis in response to pressure overload is more robust than mechanisms underlying aging and inflammation. Mechanisms are organized to present collagen content, pro- and antifibrotic gene expression, collagen cross-linking, and collagen degradation. Importantly, ventricle similarities and differences are summarized. IPAH, idiopathic pulmonary arterial hypertension; LV, left ventricle; PH, pulmonary hypertension; POSTN, periostin; RV, right ventricle; SScPAH, systemic sclerosis-associated pulmonary arterial hypertension.

Fibrosis in RV pressure overload.

Fibrosis is a common feature of HF, particularly in response to pressure overload. LV pressure overload induces interstitial and perivascular fibrosis that accompanies several forms of HF including hypertensive heart disease (26). In humans, hypertension leads to excess type I collagen over the more elastic type III (26) and as discussed above, collagen cross-linking is tightly associated with LV stiffness (36). Preclinically, pressure overload is surgically modeled by transverse aortic constriction (TAC). TAC induces fibrotic remodeling, including fibroblast activation and collagen deposition alongside upregulation of TGF-β and its downstream targets (49). Similar results have been seen in spontaneously hypertensive rats, which have elevated collagen I:III ratios in the LV (51), together linking LV afterload, fibrotic remodeling, and impaired LV function.

One of the primary distinguishing pathophysiological factors between the RV and LV is the poor tolerance of the RV to increased afterload. When pulmonary artery pressure (PAP) increases such as with PH, the RV can experience a fivefold rise in afterload compared with the ∼50% increase that occurs in the LV during systemic hypertension. Even small changes in pulmonary vascular resistance or PAP can reduce RV contractile performance and result in interstitial and perivascular fibrosis. Models of RV afterload include Sugen hypoxia (SuHx), monocrotaline (MCT), chronic hypoxia exposure, and pulmonary artery banding (PAB). While all these models increase PAP and concomitant RV dysfunction, the degree to which they result in RV failure varies, as does the fibrotic response. Adult male rats exposed to MCT had a higher percent area of RV collagen compared with controls by Picrosirius red staining (50). SuHx also resulted in higher collagen content and expression of procollagen isoforms Col1a and Col3 (52). PAB in male rabbits (53) and mice (54) increased collagen deposition. PAB not only stimulates collagen synthesis but also decreases collagen degradation, one of the very few publications to measure the degradation side of the ECM balance (55). Chronic hypoxia-induced PH appears to mildly stimulate fibrotic remodeling in the RV, with higher expression of procollagen-1 and total collagen compared with young control rats (56) and male mice (77). Some effort has been made to compare models of RV afterload and the fibrotic response, with a study in young male rats comparing SuHx, MCT, and PAB. The authors found significant but comparable RV remodeling in all models, including similar RV collagen content (78).

Given the poor capacity of the RV to tolerate pressure overload, it is tempting to hypothesize that given a similar increase in afterload that the RV would respond more robustly than the LV. Indeed, a comparison of PAB to TAC reported that the RV underwent exacerbated remodeling in response to PAB than the LV in response to TAC. PAB RV weight was almost double that of shams, whereas LV + septum weights in response to TAC only increased by ∼70%. Although the authors did not quantify fibrosis, they did show that transcriptional changes following PAB were different from TAC. Among the top regulated genes in PAB compared with TAC were ECM proteins such as POSTN and Col8a as well as other procollagens (79). Data from our laboratory also support the notion that the RV responds more robustly to pressure overload than the LV with respect to fibrosis (Fig. 2). Mice subjected to PAB upregulated expression of Col3 and POSTN compared with sham controls, supporting the observation that POSTN is consistently upregulated in RV disease (60, 61). Together, though both ventricles undergo fibrotic remodeling in response to pressure overload, the magnitude of the remodeling and perhaps the mechanisms differ, highlighting the need for future investigation and comparisons to understand RV and LV fibrotic pathways.

Figure 2.

Fibrotic gene expression in response to pressure overload (PO) in the left ventricle [LV; transaortic constriction (TAC)] and right ventricle [RV; pulmonary artery banding (PAB)]. A: periostin (POSTN) expression was higher in the PO RV and unchanged in the PO LV. B: α-smooth muscle actin (α-SMA) expression was unchanged in either PO model. C and D: expression of Col1a1 (C) and Col3 (D) were higher in the PO RV and unchanged in the PO LV compared with respective shams. E: expression of matrix metalloproteinase 2 (MMP2) was lower in the PO LV compared with sham and was unchanged in the RV; n = 4 per group. *P < 0.05, PO vs. sham within ventricle. #P < 0.1, PO vs. sham within ventricle. Gene expression was quantified by quantitative reverse transcription polymerase chain reaction (qRT-PCR) in ∼4-mo-old male C57Bl6/J mice. White bars, sham; gray bars, PO. Data are expressed as means ± SE. Adult (∼4–6 mo) C57/BL/6J male mice were subjected to either PAB (80) or TAC (81) surgery. Sham mice underwent anesthesia and thoracotomy without suturing the pulmonary artery or the aorta. Four weeks after surgery, the RV was dissected from the LV (PAB) and the LV + septum from the RV (TAC) and ventricles were flash frozen. Previous reports have not demonstrated significant effects of TAC or PAB on the opposite ventricle over this time course (80). A limitation of these experiments is that we have no direct measure of the degree of pressure overload that is induced by either PAB or TAC. qRT-PCR was performed as previously described (82).

Fibrosis and RV inflammation.

Dysregulation of inflammation is tightly linked to fibrosis in the LV. Activation of inflammatory signaling is noted in myocardial infarction (83) and ischemic cardiomyopathy (84) and results in initiation of fibrosis signaling cascades. The dynamic interplay between LV inflammation and fibrosis is well-studied and has been the focus of several reviews (85, 86). Briefly, inflammation is a primary regulator of cardiac reparative responses. Timing and regulation of inflammatory responses are critical to respond to cardiac injury. Both ventricles have resident populations of mast cells and macrophages that help monitor and regulate homeostasis (62). Acutely, after injury, these cells initiate signaling cascades to clear dysfunctional or apoptotic cardiomyocytes and other cell debris. Death or injury of cardiomyocytes stimulates the release of additional cytokines and signaling factors that further potentiate immune cell migration and myofibroblast conversion (63). Within days of injury, resident fibroblasts begin proliferating and undergo activation of myofibroblasts, resulting in ECM production (87). Myofibroblasts continue to deposit ECM proteins to maintain structural integrity of the LV in the form of replacement or reactive interstitial fibrosis. Indeed, the intensity and persistence of inflammation dictates the magnitude and duration of inflammatory and fibrotic responses. Inhibition of this inflammatory process is harmful, at least in the acute setting. In a mouse cryoinjury model, depletion of macrophage activity reduced myocardial reparative capacity because of diminished myofibroblast infiltration and fibrosis, resulting in LV dilatation and wall thinning and high mortality (64). Other studies also link LV immune cells to fibrosis. In a model of ischemic infarct in mice, monocyte chemoattractant protein-1/CCL2-deficient mice also had diminished myofibroblast accumulation compared with wild-type animals (65). Reciprocally, transgenic overexpression of the cytokine TNF-α increased collagen synthesis, deposition, and denaturation in mice (66). These mechanistic studies directly link inflammation to fibrotic remodeling in the LV.

Mechanisms of RV inflammatory-fibrotic remodeling have largely been assumed based on LV literature. However, differences exist between the RV and LV with respect to their inflammatory response. The RV, for example, possesses a greater percentage of immune cells such as macrophages and dendritic cells compared with the LV (16). Whether these homeostatic differences translate to different inflammatory activation and concomitant fibrotic remodeling in the RV is less clear. Interleukin-6 (IL-6) and C-reactive protein are independently associated with RV morphology in patients free of cardiovascular disease, independent of LV structure and function (67). High levels of circulating cytokines (IL-1β, IL-2, IL-6, and TNF-α) in patients with PH demonstrate by Kaplan–Meier mortality analysis that interleukin concentrations predict survival (68) in these patients. Overbeek et al. compared inflammatory and fibrosis pathways in two different classes of PH: systemic sclerosis-associated pulmonary arterial hypertension and IPAH. Systemic sclerosis-associated pulmonary arterial hypertension is more associated with inflammation than IPAH, presenting an opportunity to potentially unravel RV inflammation and fibrosis. The authors found that RV from patients with systemic sclerosis-associated pulmonary arterial hypertension and IPAH had high levels of inflammatory cells such as neutrophilic granulocytes, macrophages, and lymphocytes compared with healthy controls. However, despite denser inflammatory infiltrates in systemic sclerosis-associated pulmonary arterial hypertension RV, there were no differences in percent collagen area between systemic sclerosis-associated pulmonary arterial hypertension or IPAH or even healthy controls (69). Although more work needs to be done, these data preliminarily suggest that in the RV, the presence of inflammatory cells and mediators does not necessarily result in fibrotic remodeling. Though the available data linking RV inflammation and fibrosis are sparse, there is enough to suggest that the RV and LV not only respond to or possess baseline immune function, but may also differently couple inflammation with fibrotic remodeling.

Fibrosis in the aging RV and LV.

The myocardium undergoes a number of structural and functional changes as it ages. Among these changes are increased ventricular stiffness and diastolic dysfunction (88). Therefore, cardiac fibrosis has emerged as a significant contributor to age-related heart disease. Elevated collagen content and collagen cross-linking with aging render collagen stiffer and contribute to ventricular thickness and tension (72). The age-related fibrotic process impairs ventricular function and reserve, resulting in impaired diastolic function (73). A consistent increase in collagen with age has been noted across species. Furthermore, a shift from collagen III to I contributes to LV stiffness since type III is more distensible (70). Transcriptionally, expression of profibrotic factors such as collagens and fibronectin are unchanged or decrease with aging in the LV (71), pointing to posttranscriptional mechanisms of ECM deposition. Indeed, collagen cross-linking increases with age in the LV and contributes to LV stiffness (74). Furthermore, expression of antifibrotic genes gradually decreases with age (75), supporting the notion that accumulation of fibrous connective tissue in cardiac ECM is caused by slower removal of ECM components rather than increased deposition of fibrotic proteins into the matrix.

Unlike the LV, the changes in RV function with age are still contentious. Although accepted that PAP increases linearly with age (89), the resultant changes in RV function are unclear, with some groups reporting that RV systolic function is preserved (90) with others suggesting that RV systolic function declines with advanced age (91). RV diastolic function likely declines with age, as evidenced by a larger contribution of right atrial filling, suggesting altered compliance of the RV (92, 93). Similar to the LV, most reports suggest that fibrotic content as quantified by histology or hydroxyproline, increases with age in the RV (data from our laboratory, Fig. 3, A and B) (22, 23). Consistent with net deposition of collagen, in a mouse model of accelerated aging by deletion of the mitochondrial DNA polymerase, RNAseq analyses demonstrate activation of fibrotic pathways in the RV (95). Some previous work suggests that age-related increases in collagen as a percentage of total protein occur to a larger extent in the RV than the LV (76) though this distinction largely remains without consensus. What does appear to be different, however, is that the mechanisms of collagen deposition appear to differ between the ventricles, as we show that upregulation of expression of α-SMA, Col1a1, and Col3 occurs with aging in the RV but not in the LV (Fig. 3, C–G). Beyond these reports of changes in global gene expression and fibrosis, little is known about how fibrosis occurs in the aging RV. While collagen cross-linking in the LV is elevated with age, it has not been reported to differ in the aged RV compared with young in male rats (23). We were unable to find any reports that compared RV and LV pro- or antifibrotic mechanisms with age, nor did we find publications of in vitro experiments investigating aged RV fibroblasts. Given the rapidly aging US and global population and that aging is the primary risk factor for heart disease, more work is needed to understand fibrotic mechanisms of LV and RV aging.

Figure 3.

Fibrosis and fibrotic gene expression in adult (4 mo) and aged (18 mo) male and female C5lBl6/J mice. A and B: representative Picrosirius red collagen histology of adult and aged right ventricle (RV) and left ventricle (LV) (A), demonstrating higher collagen content in the aged RV and LV compared with adult matched ventricles (B). ^Higher percent area of collagen in the LV compared with the RV. C: periostin (POSTN) expression was unchanged in either ventricle with age. α-Smooth muscle actin (α-SMA; D), Col1a1 (E), and Col3 (F) expression were higher in the aged RV compared with adult but unchanged in the LV (interaction age × ventricle, P = 0.07). E: Col1a1 expression was more robustly upregulated in the RV with age compared with the LV (interaction, age × ventricle). G: expression of matrix metalloproteinase 2 (MMP2) was lower in the aged LV compared with adult but was unchanged with age in the RV. *P < 0.05, adult vs. aged within ventricle. ^P < 0.05 main effect of ventricle (LV > RV). $P < 0.05 interaction age × ventricle. Picrosirius red staining was performed as previously described (94). Gene expression was quantified by quantitative reverse transcription polymerase chain reaction (qRT-PCR) (82). White bars, adult; gray bars, aged. Data are expressed as means ± SE.

EVIDENCE FOR RV-CENTRIC EFFECTS OF PUTATIVE ANTIFIBROTIC THERAPIES

If fibrosis is casually linked to impaired cardiac function, then identification of therapies that attenuate, or better yet, reverse fibrosis should have profound impact on patient outcomes. However, because fibrosis is not innately maladaptive, discovering antifibrotic therapies will be nuanced and will require considerable finesse to develop without adverse outcomes in the ventricles. Though to our knowledge the balance of beneficial and fibrotic remodeling has not yet been studied in the RV, some ECM protein deposition is likely needed to provide mechanical support and prevent excessive dilatation in the RV. Given the chronic and temporal nature of HF and age-associated cardiac fibrosis, it is also reasonable to assume that successful antifibrotic interventions would need to reverse existing fibrosis that is composed of cross-linked collagen resistant to breakdown. Because of the persistent nature of collagen cross-linking, antifibrotic therapies have largely focused on the inhibition of collagen synthesis or inhibition of collagen cross-linking mechanisms, not on the reversal of these processes. Successful implementation of antifibrotic therapies will likely require integration of synthesis, cross-linking, and degradation pathways for therapeutic success.

An indirect approach to attenuate fibrotic remodeling is through resolution of cardiac stressors. For example, the resolution of high afterload may allow the heart to “catch up” and abrogate deposited collagen. In support of this notion, the ACE inhibitor lisinopril has antifibrotic effects (decreased hydroxyproline and collagen volume fraction) in humans with hypertensive heart disease (96). The angiotensin receptor blocker losartan lowered late gadolinium enhancement compared with placebo in patients with hypertrophic cardiomyopathy (97). The anti-inflammatory effects of statins have been investigated as antifibrotic therapies with rosuvastatin attenuating cardiac fibrosis in hypertensive rats (98) and patients with systolic HF (99). Mechanical unloading using LV assist device, however, has shown conflicting reports, reversing LV fibrosis (100) and increasing insoluble collagen concentrations (101), perhaps suggesting that simple unloading is not sufficient, but rather interventions must also directly reverse remodel the myocardium. In the RV, PAB causes robust collagen deposition that in some cases can be reversed by debanding (102). However, other reports show that debanding the pulmonary artery does not improve collagen content (53) despite improvements in RV function and resolution of hypertrophy. Prostacyclins are a cornerstone of PH treatment, and recent interest has developed in their direct effect on the RV. A study using both SuHx and PAB in rats investigated whether the prostacyclin iloprost could reverse RV fibrosis. Iloprost robustly decreased collagen content alongside lower expression of procollagen and connective tissue growth factor expression. Iloprost decreased cardiac myofibroblast activation and migration and attenuated MMP-9 gene expression as well as other autophagic genes associated with collagen degradation (52). Importantly, many of these improvements in RV fibrotic remodeling occurred independent of changes in pulmonary artery pressure, again suggesting that successful antifibrotic therapies will likely require more than just attenuating overload and must provide direct cardioprotective effects.

Exercise is potently cardioprotective and has been suggested as a potential antifibrotic therapy. Exercise training attenuated LV collagen deposition with advanced age as well as the age-related decline in MMP1, MMP2, and MMP14 activity and expression (103). Exercise also attenuated LV fibrosis by inhibiting advanced glycation end product accumulation (44). However, exercise did not attenuate age-associated increases in collagen content in the RV, despite significantly lowering percent collagen area in the LV and septum (23), further supporting the notion that the ECM is regulated differently in each ventricle with age. Though exercise has been historically avoided in patients with elevated pulmonary pressures and right-sided disease because of fear of adverse events, limited preclinical data suggest a benefit of exercise training as antifibrotic therapy in experimental PH (104). With the growing recognition that exercise is safe and beneficial for these patients, we hope to gain a more mechanistic understanding of how exercise may be antifibrotic in the RV.

CONCLUSIONS AND REMAINING QUESTIONS

Throughout this review, where possible, we have included details of animal/subject sex. Biological sex clearly contributes to differential outcomes and mechanisms of cardiac remodeling including fibrosis. An entire review could be written on the topic of sex differences in fibrosis in the right (and left) ventricles, which is unfortunately beyond the scope of the current work. However, we wanted to acknowledge the significance of sex differences with respect to our conclusions. LV disease and aging are characterized by sex differences in fibrotic outcomes. Our group has previously published that age-related alterations in cardiac ECM in the LV are sex dependent, both with respect to quantity of fibrotic content and mechanisms of deposition (75). Furthermore, in response to pressure overload via TAC (105) or isoproterenol (106), male mice undergo more pronounced fibrotic remodeling than females. Interestingly, in the latter report using isoproterenol, differences in the cardiac fibroblasts themselves in terms of activation status and expression of maladaptive genes were responsible for the sex differences in fibrosis, rather than sex hormones. RV failure has long been recognized to present with sex differences. Females more frequently develop PH but fare better postdiagnosis than male counterparts. Indeed, male sex is associated with reduced survival and exacerbated RV fibrosis in preclinical models such as PAB (57). Although estrogen is undoubtedly cardioprotective and attenuates RV fibrosis and expression of profibrotic genes (58, 59), surgical depletion of ovarian estrogen does not always normalize fibrotic remodeling in the diseased RV (57) suggesting the involvement of mechanisms beyond sex hormones. Mechanistic understanding of the development and hopefully regression of fibrosis in the RV will necessitate the inclusion of male and female models.

The RV and LV are embryologically, anatomically, and physiologically distinct. At birth, changes in pulmonary and systemic pressure likely result in developmental changes and different cellular composition of the two ventricles, resulting in a higher composition of immune cells in the RV and yet unclear differences in collagen and ECM content. Given these differences, it should not be surprising that the two ventricles also respond to pathological load via distinct mechanisms. These differences are evident in the ventricle-specific mechanisms of fibrosis. In response to pressure overload, inflammation, and with advanced age, the RV becomes fibrotic. It deposits more collagen likely because of elevated profibrotic signaling but also probably because of diminished collagen breakdown. However, posttranslational modification of ECM proteins (lysyl oxidase, advanced glycation end products) has not yet been elucidated in the RV. While the RV fibroblast is activated to the pathogenic myofibroblast, specific mechanisms of activation still are unclear due to a paucity of mechanistic in vitro studies in isolated RV fibroblasts. POSTN is consistently and robustly upregulated in RV disease but the significance of this molecule in RV disease has not yet been tested. These RV-specific mechanisms and gaps in current understanding are summarized in Fig. 4. Importantly, however, the direct impact of fibrotic remodeling on RV function remains unclear despite the predictive power of RV function on patient survival. Future work must aim to understand ventricle- and likely disease-specific mechanisms of RV fibrotic remodeling. Advanced understanding of mechanisms of RV fibrosis such as myofibroblast differentiation and ECM degradation pathways including collagen cross-linking is necessary for the pursuit of antifibrotic therapeutics and improved RV health.

Figure 4.

Summary of current understanding of right ventricle (RV) fibrosis. The healthy RV extracellular matrix (ECM) is under constant regulation to balance collagen synthesis and degradation to maintain RV function and integrity. However, cardiac stress from pathological conditions such as pressure overload, inflammation, or aging alters cell composition and extensively modifies the ECM to accumulate myofibroblasts and ECM proteins. Periostin is consistently upregulated in RV disease and likely plays an important role in RV remodeling. Myofibroblast differentiation, matrix metalloproteinase (MMP) activity, and collagen degradation or collagen cross-linking remain poorly understood in the fibrotic RV.

GRANTS

This work was supported by National Institute on Aging Grant K01 AG058810 (to D.R.B.) and American Heart Association Grant 897622 (to B.D.M.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

D.R.B. conceived and designed research; B.D.M. and S.K.S. performed experiments; B.D.M. analyzed data; B.D.M. interpreted results of experiments; B.D.M., S.K.S., and D.R.B. prepared figures; B.D.M., S.K.S., and D.R.B. drafted manuscript; B.D.M., S.K.S., and D.R.B. edited and revised manuscript; B.D.M., S.K.S., and D.R.B. approved final version of manuscript.