Circulating Biomarkers of Myocardial Fibrosis
A number of molecules, detectable in either the serum or plasma in humans by using immunoassay methods, have recently been proposed as biomarkers of myocardial fibrosis (Table 1). However, in most cases, demonstration of an association between the biomarker and histologically assessed myocardial fibrosis (i.e., the blood level of the biomarker directly correlates with either CVF or CIVF or CIIIVF, and after a therapeutic intervention, it changes in parallel with the changes in some of these 3 histological parameters) is lacking or remains inconclusive. Some examples may serve to illustrate the importance of this initial validation of the candidate molecules.
As presented in Table 1, among the many circulating molecules proposed as biomarkers of myocardial fibrosis in humans, only 2 collagen-derived serum peptides have been shown to be associated with myocardial fibrosis: the carboxy-terminal propeptide of procollagen type I (PICP), formed during the extracellular conversion of procollagen type I into mature fibril-forming collagen type I by the enzyme procollagen type I carboxy-terminal proteinase, and the amino-terminal propeptide of procollagen type III (PIIINP), formed during the extracellular conversion of procollagen type III into mature fibril-forming collagen type III by the enzyme procollagen type III amino-terminal proteinase (Figure 2). This topic has been reviewed previously by Prockop and Kivirikko.
(Enlarge Image)
Figure 2.
Extracellular Generation and Release Into the Bloodstream of Collagen-Derived Serum Peptides
(A) The amino-terminal propeptide of procollagen type I (PINP) and the carboxy-terminal propeptide of procollagen type I (PICP) originate during the conversion of procollagen type I in collagen type I, and (B) the amino-terminal propeptide of procollagen type III (PIIINP) and the carboxy-terminal propeptide of procollagen type III (PIIICP) originate during the conversion of procollagen type III in collagen type III. (1) Since PICP is formed in a 1:1 stoichiometric ratio to the collagen type I molecule formed, the propeptide concentrations in tissue fluids and blood are direct indicators of ongoing collagen type I synthesis. (2) The cleavage at the amino-terminus proceeds at a relatively slow rate and, thus, partially processed procollagen molecules (known as pN-collagen type III molecules) are found on the surface of collagen type III fibers. Therefore, the PIIINP concentrations in tissue fluids and blood are not direct indicators of ongoing collagen type III synthesis. (3) Although PIIICP is set free, it is unknown whether it reaches the bloodstream. PICP = procollagen type I carboxy-terminal proteinase; PINP = procollagen type I amino-terminal proteinase; PIIICP = procollagen type III carboxy-terminal proteinase; PIIINP = procollagen type III amino-terminal proteinase.
Serum PICP levels have been found to be highly correlated with total CVF in patients with hypertensive heart disease (HHD) without and with HF. In addition, it has been shown that there is a robust correlation between serum PICP levels and CIVF in patients with HHD and HF. Of interest, serum PICP levels and CVF reportedly changed in parallel in response to losartan therapy in patients with HHD without HF and in response to torasemide therapy in patients with HHD and HF. These levels also varied similarly in HF patients with idiopathic dilated cardiomyopathy (DCM) treated with spironolactone.
Serum PIIINP has been found to be highly correlated with CIIIVF in HF patients with ischemic heart disease or idiopathic DCM. In addition, the reduction in the extent of CVF observed in HF patients with idiopathic DCM treated with spironolactone was accompanied by a significant reduction in serum PIIINP.
Although the small carboxy-terminal telopeptide released by the action of matrix metalloproteinase (MMP)-1 on collagen type I fibers (CITP) may reflect collagen degradation, the information regarding its relation to myocardial fibrosis is controversial. A direct correlation of CITP with CIVF was reported in HF patients with ischemic heart disease or idiopathic DCM; however, Izawa et al. found that serum CITP levels were lower in HF patients with idiopathic DCM and severe myocardial fibrosis than in patients with mild to moderate fibrosis. Furthermore, the reduction in fibrosis with spironolactone in patients with severe fibrosis was accompanied by an increase in serum CITP.
It has been shown experimentally that microRNA-21 regulates the activity of cardiac fibroblasts and participates in the development of myocardial fibrosis. Interestingly, the circulating levels of microRNA-21 were directly correlated with myocardial expression of collagen type I messenger ribonucleic acid in patients with isolated severe aortic stenosis undergoing aortic valve replacement surgery. However, no data were provided in this study regarding CVF and CIVF or their potential association with microRNA-21.
Although the available evidence suggests that the small lectin-like protein galectin-3 mediates fibrosis in different experimental models of HF, no associations were found between plasma galectin-3 levels and CVF, CIVF, and CIIIVF in patients with HHD and HF. Conversely, even though CVF increased significantly in HF patients with idiopathic DCM 6 months after left ventricular assist device (LVAD) support, plasma galectin-3 decreased significantly in the same patients. Although it has been shown that the matricellular protein connective tissue growth factor is up-regulated in different models of myocardial fibrosis, its plasma level did not correlate with CVF in HF patients with nonischemic DCM, and it remained stable 6 months after LVAD support in these patients, despite the fact that CVF increased significantly.
Abundant experimental evidence supports the notion that transforming growth factor (TGF)-β plays an important role in the pathogenesis of myocardial fibrosis (reviewed by Dobaczewski et al.), and plasma TGF-β levels were found to be directly correlated with the LV myocardial expression of collagen type I messenger ribonucleic acid in patients with aortic stenosis undergoing valve replacement. However, no correlation of plasma TGF-β with either CVF or CIVF was provided in this study. In addition, it has been reported that plasma TGF-β levels were higher in control subjects than in HF patients with idiopathic DCM that had abnormally high CVF values. Furthermore, although CVF increased significantly in these patients 6 months after LVAD support, plasma TGF-β did not change.
Growth differentiation factor-15 is a member of the TGF-β cytokine superfamily that has been shown to be associated with myocardial fibrosis in a genetic model of HF. Although plasma levels of growth differentiation factor-15 reportedly correlate directly with CVF in patients with idiopathic DCM before LVAD support, the reduction in these levels observed 1 month after LVAD support was accompanied by an increase in CVF.
No correlations between CVF and serum MMP-1, serum tissue inhibitor of matrix metalloproteinases-1, and their ratio (which may serve as an indirect index of circulating unbound active MMP-1) have been found in patients with HHD and HF. However, it is noteworthy that those patients with higher values of serum MMP-1 and serum MMP-1:tissue inhibitor of matrix metalloproteinases-1 ratio exhibited reduced levels of perimysial and endomysial collagen. This finding suggests that these molecules may be related more to the loss of the physiological mysial collagen scaffold than to the accumulation of pathologic nonmysial collagen.
Experimental findings support the notion that the matricellular protein osteopontin seems to mediate fibrogenic actions in HF models. However, plasma concentrations of osteopontin are reportedly not associated with CVF or CIVF in patients with HHD and HF. Furthermore, in HF patients with nonischemic DCM, plasma osteopontin levels tended to decrease whereas CVF increased significantly during LVAD support.
Cardiotrophin-1 is a member of the interleukin-6 superfamily that behaves in vitro as a profibrotic cytokine in cultured animal and human cardiac fibroblasts, and in vivo it induces cardiac fibrosis in rats. Recently, an association between the myocardial expression of cardiotrophin-1 and collagen types I and III has been reported in the fibrotic myocardium of patients with HF of hypertensive origin. However, the plasma levels of cardiotrophin-1 were not correlated with CVF, CIVF, or CIIIVF in the same patients.
Validation of the Candidate Biomarkers: Available Clinical Evidence
A number of molecules, detectable in either the serum or plasma in humans by using immunoassay methods, have recently been proposed as biomarkers of myocardial fibrosis (Table 1). However, in most cases, demonstration of an association between the biomarker and histologically assessed myocardial fibrosis (i.e., the blood level of the biomarker directly correlates with either CVF or CIVF or CIIIVF, and after a therapeutic intervention, it changes in parallel with the changes in some of these 3 histological parameters) is lacking or remains inconclusive. Some examples may serve to illustrate the importance of this initial validation of the candidate molecules.
Biomarkers With Proven Evidence of Their Association With Myocardial Fibrosis
As presented in Table 1, among the many circulating molecules proposed as biomarkers of myocardial fibrosis in humans, only 2 collagen-derived serum peptides have been shown to be associated with myocardial fibrosis: the carboxy-terminal propeptide of procollagen type I (PICP), formed during the extracellular conversion of procollagen type I into mature fibril-forming collagen type I by the enzyme procollagen type I carboxy-terminal proteinase, and the amino-terminal propeptide of procollagen type III (PIIINP), formed during the extracellular conversion of procollagen type III into mature fibril-forming collagen type III by the enzyme procollagen type III amino-terminal proteinase (Figure 2). This topic has been reviewed previously by Prockop and Kivirikko.
(Enlarge Image)
Figure 2.
Extracellular Generation and Release Into the Bloodstream of Collagen-Derived Serum Peptides
(A) The amino-terminal propeptide of procollagen type I (PINP) and the carboxy-terminal propeptide of procollagen type I (PICP) originate during the conversion of procollagen type I in collagen type I, and (B) the amino-terminal propeptide of procollagen type III (PIIINP) and the carboxy-terminal propeptide of procollagen type III (PIIICP) originate during the conversion of procollagen type III in collagen type III. (1) Since PICP is formed in a 1:1 stoichiometric ratio to the collagen type I molecule formed, the propeptide concentrations in tissue fluids and blood are direct indicators of ongoing collagen type I synthesis. (2) The cleavage at the amino-terminus proceeds at a relatively slow rate and, thus, partially processed procollagen molecules (known as pN-collagen type III molecules) are found on the surface of collagen type III fibers. Therefore, the PIIINP concentrations in tissue fluids and blood are not direct indicators of ongoing collagen type III synthesis. (3) Although PIIICP is set free, it is unknown whether it reaches the bloodstream. PICP = procollagen type I carboxy-terminal proteinase; PINP = procollagen type I amino-terminal proteinase; PIIICP = procollagen type III carboxy-terminal proteinase; PIIINP = procollagen type III amino-terminal proteinase.
Serum PICP levels have been found to be highly correlated with total CVF in patients with hypertensive heart disease (HHD) without and with HF. In addition, it has been shown that there is a robust correlation between serum PICP levels and CIVF in patients with HHD and HF. Of interest, serum PICP levels and CVF reportedly changed in parallel in response to losartan therapy in patients with HHD without HF and in response to torasemide therapy in patients with HHD and HF. These levels also varied similarly in HF patients with idiopathic dilated cardiomyopathy (DCM) treated with spironolactone.
Serum PIIINP has been found to be highly correlated with CIIIVF in HF patients with ischemic heart disease or idiopathic DCM. In addition, the reduction in the extent of CVF observed in HF patients with idiopathic DCM treated with spironolactone was accompanied by a significant reduction in serum PIIINP.
Biomarkers With Inconclusive Evidence of Their Association With Myocardial Fibrosis
Although the small carboxy-terminal telopeptide released by the action of matrix metalloproteinase (MMP)-1 on collagen type I fibers (CITP) may reflect collagen degradation, the information regarding its relation to myocardial fibrosis is controversial. A direct correlation of CITP with CIVF was reported in HF patients with ischemic heart disease or idiopathic DCM; however, Izawa et al. found that serum CITP levels were lower in HF patients with idiopathic DCM and severe myocardial fibrosis than in patients with mild to moderate fibrosis. Furthermore, the reduction in fibrosis with spironolactone in patients with severe fibrosis was accompanied by an increase in serum CITP.
It has been shown experimentally that microRNA-21 regulates the activity of cardiac fibroblasts and participates in the development of myocardial fibrosis. Interestingly, the circulating levels of microRNA-21 were directly correlated with myocardial expression of collagen type I messenger ribonucleic acid in patients with isolated severe aortic stenosis undergoing aortic valve replacement surgery. However, no data were provided in this study regarding CVF and CIVF or their potential association with microRNA-21.
Although the available evidence suggests that the small lectin-like protein galectin-3 mediates fibrosis in different experimental models of HF, no associations were found between plasma galectin-3 levels and CVF, CIVF, and CIIIVF in patients with HHD and HF. Conversely, even though CVF increased significantly in HF patients with idiopathic DCM 6 months after left ventricular assist device (LVAD) support, plasma galectin-3 decreased significantly in the same patients. Although it has been shown that the matricellular protein connective tissue growth factor is up-regulated in different models of myocardial fibrosis, its plasma level did not correlate with CVF in HF patients with nonischemic DCM, and it remained stable 6 months after LVAD support in these patients, despite the fact that CVF increased significantly.
Abundant experimental evidence supports the notion that transforming growth factor (TGF)-β plays an important role in the pathogenesis of myocardial fibrosis (reviewed by Dobaczewski et al.), and plasma TGF-β levels were found to be directly correlated with the LV myocardial expression of collagen type I messenger ribonucleic acid in patients with aortic stenosis undergoing valve replacement. However, no correlation of plasma TGF-β with either CVF or CIVF was provided in this study. In addition, it has been reported that plasma TGF-β levels were higher in control subjects than in HF patients with idiopathic DCM that had abnormally high CVF values. Furthermore, although CVF increased significantly in these patients 6 months after LVAD support, plasma TGF-β did not change.
Growth differentiation factor-15 is a member of the TGF-β cytokine superfamily that has been shown to be associated with myocardial fibrosis in a genetic model of HF. Although plasma levels of growth differentiation factor-15 reportedly correlate directly with CVF in patients with idiopathic DCM before LVAD support, the reduction in these levels observed 1 month after LVAD support was accompanied by an increase in CVF.
Biomarkers With Lack of Evidence of Their Association With Myocardial Fibrosis
No correlations between CVF and serum MMP-1, serum tissue inhibitor of matrix metalloproteinases-1, and their ratio (which may serve as an indirect index of circulating unbound active MMP-1) have been found in patients with HHD and HF. However, it is noteworthy that those patients with higher values of serum MMP-1 and serum MMP-1:tissue inhibitor of matrix metalloproteinases-1 ratio exhibited reduced levels of perimysial and endomysial collagen. This finding suggests that these molecules may be related more to the loss of the physiological mysial collagen scaffold than to the accumulation of pathologic nonmysial collagen.
Experimental findings support the notion that the matricellular protein osteopontin seems to mediate fibrogenic actions in HF models. However, plasma concentrations of osteopontin are reportedly not associated with CVF or CIVF in patients with HHD and HF. Furthermore, in HF patients with nonischemic DCM, plasma osteopontin levels tended to decrease whereas CVF increased significantly during LVAD support.
Cardiotrophin-1 is a member of the interleukin-6 superfamily that behaves in vitro as a profibrotic cytokine in cultured animal and human cardiac fibroblasts, and in vivo it induces cardiac fibrosis in rats. Recently, an association between the myocardial expression of cardiotrophin-1 and collagen types I and III has been reported in the fibrotic myocardium of patients with HF of hypertensive origin. However, the plasma levels of cardiotrophin-1 were not correlated with CVF, CIVF, or CIIIVF in the same patients.