Emerging pathways of communication between the heart and non-cardiac organs

The breakthrough discovery of cardiac natriuretic peptides provided the first direct demonstration of the connection between the heart and the kidneys for the maintenance of sodium and volume homeostasis in health and disease. Yet, little is still known about how the heart and other organs cross-talk. Here, we review three physiological mechanisms of communication linking the heart to other organs through: i) cardiac natriuretic peptides, ii) the microRNA-208a/mediator complex subunit-13 axis and iii) the matrix metalloproteinase-2 (MMP-2)/C-C motif chemokine ligand-7/cardiac secreted phospholipase A2 (sPLA2) axis – a pathway which likely applies to the many cytokines, which are cleaved and regulated by MMP-2. We also suggest experimental strategies to answer still open questions on the latter pathway. In short, we review evidence showing how the cardiac secretome influences the metabolic and inflammatory status of non-cardiac organs as well as the heart.


Introduction
Diseases with strong metabolic and inflammatory components (ischemic heart disease, arthritis, neurodegeneration and cancer) are leading causes of morbidity, extremely high costs associated with health care, and mortality worldwide [1] . Given the incomplete understanding of these diseases, new studies to decipher organ-specific cross-talk between inflammation and metabolic pathways are constantly needed.
This paper aims to: (i) introduce historical background for the groundbreaking notion that the heart exerts an endocrine function [2][3][4][5] , (ii) review briefly two other recent mechanisms-the cardiac-specific micro-RNA (miR)-208a/mediator complex subunit-13 (MED13) axis [6] and the matrix metalloproteinase 2 (MMP-2)/C-C motif chemokine ligand 7 (CCL7)/ cardiac secretory phospholipase A2 (sPLA2) axis-by which the heart modulates lipid metabolism in noncardiac organs (e.g., the liver) [7][8][9] and (iii) suggest experimental approaches to elucidate the mechanism that regulates the cardiac-specific origin of sPLA2 in MMP-2 deficiency as well as delineate the biochemical pathway by which the dyad of MMP-2 and monocyte-chemotactic protein 3 (MCP-3)/CCL7 modulates cholesterol homeostasis in the heart and other organs. Outputs of this latter line of research have the potential to open up new avenues for modulating cholesterol metabolism [e.g., at the levels of MMP-2 activity or signaling mediators positioned downstream of C-C chemokine receptor type 2 (CCR2)] in inflammatory and metabolic disorders.
The heart has an endocrine function consisting of the release of natriuretic peptides Rows 1 to 8 in Table 1 show selected discoveries related to the ability of the heart atrial muscle cells of mammals to synthesize, store within specific atrial granules, and release cardiac natriuretic peptides (cNPs). The mechanism of action of cNPs is summarized in Box 1. These findings by the research team of Dr. de Bold were the first direct demonstration that the heart has an endocrine function (summarized in Table 1, rows 1 to 8). In Table 2, we summarize selected observations describing how the cNP system influences hypertrophic growth, fibrosis, and cardiac remodeling/ dysfunction. For a summary of the influences of cNPs on adipose tissue biology and metabolism, please see Table 3.
A cardiac-specific miR-208a/MED13 axis further connects the heart with other organs Recent studies [6,[50][51] have identified cardiac-specific miR-208a as a negative regulator of the subunit 13 of mediator complex MED13 (summarized in Table 1, row 10). MED13 regulates the transcription of many nuclear receptor genes involved in fatty acid oxidation as well as influencing the activity of as-yet-unidentified secreted/circulating factors, which connect the activity of cardiac MED13 with the metabolism and energy homeostasis program of non-cardiac organs, such as the liver and adipose tissue (row 10 of Table 1).

MMP-2 deficiency is associated with elevated secreted/circulating cardiac PLA2 activity
In 2015, a possibly new endocrine system was postulated, by which the heart influences cardio-hepatic lipid metabolism, hepatic sensitivity to dietary cholesterol, systemic inflammatory status, severity of fever and energy expenditure [7][8][9] . By investigating the pathophysiological consequences of MMP-2 deficiency in MMP-2 null (Mmp2 -/-) mice, it was found that MMP-2 governs the secretion of a highly pro-inflammatory cardiac-specific phospholipase A2 activity (named 'cardiac' sPLA2). This finding provides a plausible and novel mechanism that could explain, at least partially, why human MMP-2 deficiency results in pediatric inflammatory arthritis with relentless bone loss, inflammation, cardiac developmental defects and other metabolic abnormalities such as hirsutism and dwarfism [7][8][9] . Two years after the identification of the MMP-2/cardiac sPLA2 axis [7][8][9] , there are key questions which warrant further investigation including: What is the molecular identity of the MMP-2-regulated sPLA2? What determines the cardiac origin of this sPLA2 in MMP-2 deficiency? We address these questions in the sections below.
Cardiac sPLA2 may belong to the family of classical secreted phospholipases but its molecular composition is unknown Up to now unsuccessful, previous attempts to identify cardiac sPLA2 have used targeted time-resolved immunofluorescence assays (TRIFA) [8] or RT-PCR with reagents targeting the 31 different PLA2s (including classical and atypical, cytosolic and secreted enzymes) [7][8] as well as conventional mass spectrometry, which is not inherently quantitative (the authors' unpublished data). Activity inhibition studies have suggested that cardiac sPLA2 may be a mixture of indoxam-resistant (e.g., PLA2G1B, PLA2G2D, PLA2G2F, PLA2G10) and indoxam-sensitive (e.g., PLA2G2E, PLA2G5) sPLA2s or a new member of the sPLA2 family [8] .
To date, identifying the isoforms responsible for cardiac sPLA2 activity has been challenging calling for unbiased, highly sensitive and quantitative identification strategies such as a proteomics approach coupled with stable isotope-labelling with amino acids in vivo (mouse SILAC), a technique that has revolutionized the field of quantitative proteomics making it feasible to quantitate protein expression in mouse organs in two states [52][53] . Applying such a strategy (Box 2) has the added advantages of enabling the identification and quantification of all PLA2s deregulated (up-or downregulated) in MMP-2-deficiency along with any other proteomics abnormalities. These resultant proteomic signature of MMP-2 deficiency could serve as biomarker of disease activity or as new therapeutic target in patients.
MMP-2, CCL7 and organ-homing immune cells govern cardiac sPLA2 release in an organ-specific fashion  [2] The atrial extract (i) decreases blood pressure and slightly increases hematocrits; (ii) rapidly increases the concentration and urinary excretion of sodium and chloride (≥30-fold), the urine volume (~10-fold), the potassium excretion (two-fold); (iii) contains a potent natriuretic and chloriuretic factor, which strongly inhibits the renal tubular NaCI reabsorption.
ANP; Can extracts from other sources induce a natriuretic and diuretic response? (de Bold AJ and Salerno TA, 1983) [3] Natriuresis and diuresis is induced by atrial extracts from all mammalian species, frog atrial and ventricular extracts, hen ventricle extracts (only diuresis), and not by hen atrial extracts or rat tissue extracts other than the atrial extract; (ii) Natriuretic activity is restricted to heart.
ANP; What are the molecules responsible for these activities? (de Bold AJ and Flynn TG, 1983) [4] Cardionatrin I, which also has effect on vascular smooth muscle tone, has a molecular mass of 5.1 kDa by urea-SDS-PAGE and 49 amino acid residues one of which is cysteine.
ANP; What is the common precursor of Cardionatrin I and other atrial peptides? (Flynn TG et al., 1985) [5] (i) Cardionatrin IV, consisting of 126 amino acids, has a molecular mass of 19 kDa by urea-SDS-PAGE, and begins immediately after the signal peptide sequence of procardionatrin at residue 25. It does not contain residues 151 and 152, which are arginines; (ii) Cardionatrin III begins at residue 73 and Cardionatrin I begins at residue 123; (iii) Cardionatrins I, III, many cleavage fragments thereof and numerous versions of the carboxyl terminal portion of Cardionatrin I are products derived from a common precursor, Cardionatrin IV; (iv) Cardionatrins I-IV peptides are derived from preprocardionatrin, a common precursor of 152 amino acids (in the rat); (v) The biologically active sequences of the atrial natriuretic factor is contained in the COOHterminal portion of the molecule.
BNP; Identification in porcine brain of a novel natriuretic peptide (Sudoh T et al., 1988) [10] BNP contains 26 amino acid residues, two Cys residues, seven amino acid substitutions and one addition of (Arg) compared to α-ANP; (ii) BNP possesses diuretic-natriuretic (e.g., increase in urine output, Na + , K + , Cl -excretion) and hypotensive (decrease in mean blood pressure) responses similar to that of ANP; (iii) There may be a dual mechanism involving both ANP and BNP to control physiological functions such as water intake and salt appetite.
BNP; What is the intracellular localization of BNP in human cardiac myocytes? (Nakamura S et al., 1991) [11] BNP is specifically localized in only some of the secretory granules of the human atrium and ventricle that contain ANP, as shown with different patients (with aortic regurgitation, mitral regurgitation or autoptic); (ii) The atrium is the major production site of BNP; (iii) Together, ANP and BNP allows the human heart to regulate blood pressure and body fluid.
ANP and BNP; Natriuretic peptides circulate in blood (Clerico et al., 2011 and citations therein) [12] (i) Posttranslational processing of proBNP is required for secretion and bioactivity-this process is impaired in patients with heart failure leading to biologically inactive BNP; (ii) proBNP-derived fragments (e.g., the intact and glycosylated forms of precursor proBNP, NH2-terminal-truncated BNP form 3-32) circulate in human plasma in addition to bioactive BNP1-32; (iii) In plasma of patients with heart failure, a significant portion of immunoreactive B-type related peptides is comprised of intact or glycosylated forms of proBNPthis suggests that plasma proteases cleave the circulating proBNP to produce biologically active BNP; (iv) In experimental models and in patients with chronic heart failure, a resistance to ANP and BNP is observed; possible mechanisms of resistance to biological effects of ANP and BNP may operate at: a) the pre-receptor level (e.g., existence of inactive natriuretic peptides in plasma, increase in inactivation and degradation of active natriuretic peptides, decreased renal filtration), b) receptor level (e.g., downregulation of NPR-A and NPR-B in target tissues, altered ANP/BNP receptor binding or desensitization, or c) post-receptor level (e.g., altered intracellular signaling).
CNP; Are cardiomyocytes able to produce CNP? (Del Ry S et al., 2011) [13] (i) Both HUVEC and H9c2 muscle cells express CNP (150 and 200 bp); which can be confirmed in neonatal rat primary cardiomyocytes; (ii) CNP can be immunodetected in both H9c2 cells (by radioimmunologic assay) and cardiomyocytes of pig hearts; (iii) CNP is constitutively expressed in cardiomyocytes. Table 1 Discovery of cardiac natriuretic peptides and the heart-specific miR-208a/MED13 axis (continued)

Molecules investigated; Research question; Authors
Author's main results and conclusions ANP and BNP; Biological factors and pathophysiological mechanisms that stimulate the production /release of natriuretic peptides (Clerico et al., 2011 and citations therein) [12] (i) The production/release of cNPs is stimulated by: a) Ang II, ET1, influx), f) thyroid hormones (through thyroid hormone regulatory element), g) corticosteroids (through glucocorticoid responsive element), and h) estrogens; (ii) The production and release of BNP from ventricular cardiomyocytes is stimulated by inflammation, ventricular hypertrophy, and fibrosis; (iii) Even in isolated and cultured ventricular cells, myocardial ischemia can induce the synthesis/secretion of BNP and its related peptides; (iv) Both ANP and BNP transcription may be activated by the hypoxia-inducible factor-1a (which is induced under low oxygen conditions).
Cardiac specific miR-208a and MED13; How does cardiac MED13 influence whole body metabolism? (Grueter CE et al., 2012) [6] (i) Pharmacologic inhibition of the cardiac-specific miR-208a confers resistance to diet-induced obesity (e.g., smaller visceral WAT and subscapular BAT, normal glucose response, lower fasting insulin levels) with beneficial metabolic effects (e.g., reduced serum triglyceride and cholesterol levels); (ii) miR-208a is a negative transcriptional regulator of MED13 in the heart. Among the functions of MED13 are: (a) to inhibit expression of metabolic genes regulated by NRs (e.g., Gpd2, Thrsp, Cidea, Elovl6, Eno1, PPARγ Tkt), (b) to control whole-body metabolic homeostasis (e.g., αMHC-Med13 TG mice, with increased cardiac expression of MED13, show enhanced metabolic rate, diminished serum triglyceride and cholesterol levels, resistance to diet-induced obesity including less fat mass versus WT littermates, reduced visceral WAT and subscapular BAT mass as well as less adipocyte size and less lipid accumulation, improved glucose (tolerance) response, lowered plasma lipid levels, and improved whole-body insulin sensitivity, and (c) regulate energy expenditure (e.g., increased oxygen consumption, carbon dioxide production) in mice; (iii) MED13 deficiency in the heart increases susceptibility to metabolic syndrome and diet-induced obesity in mice, as shown with Med13 cardiac knockout mice versus Med13fl/fl littermates on HF diet. (iv) Circulating factors may relay MED13 activity from the heart to other organs but these factors remain elusive. Abbreviations
The role of NPRA/cGMP signaling in regulation of MMPs and other factors (e.g., proinflammatory mediators) (Vellaichamy et al., 2005) [45] (i) There is a link between Npr1 gene disruption in mice and the expression and activation of matrix metalloproteinases (e.g., MMP-2, MMP-9) and pro-inflammatory cytokines that play critical roles in cardiac hypertrophy, fibrosis, and extra cellular matrix remodeling; (ii) Hearts from Npr1 -/homozygous mice at an early age versus age-matched WT (Npr1 +/+ ) control mouse hearts show strongly activated genes (e.g., MMP-2 and MMP-9 by 3-5-fold, TNF-α by 8-fold) -these genes remain activated in adult mice; (iii) In Npr1 -/mice treated with GM6001, a MMP inhibitor, the activities of MMP-9 and MMP-2 are decreased (between 3 and 5 fold), fibrosis is reduced (75%), ventricular dilatation is attenuated, and fractional shortening is improved. These observations implicate MMPs in myocardial fibrosis and cardiac hypertrophy.
Abbreviations: Nppa: pro-ANP gene; TIMP: tissue inhibitor of metalloproteinase; Npr1: NPR-A gene. Table 3 Effects of the cardiac natriuretic peptide system in the cardio-adipose axis

Selected observations
The role of the cardiac natriuretic peptide system in adipose tissue biology and metabolism (Sarzani et al., 2008 and citations therein) [47] [48] (i) There is an inverse correlation between BMI (e.g.,≥30 kg/m2) and plasma BNP and NH2-terminal-proBNP values, both in healthy subjects and in patients with heart failure; (ii) The increased expression of NPR-C in the adipose tissue may lead to the peripheral clearance of BNP; nevertheless, this increase in NPR-C should have no influence in modulating the NH2terminal-proBNP levels in obese subjects. (iii) Some alterations of peripheral degradation of ANP and BNP in individuals having the corin I555 (P568) allele; (iv) The production and secretion of BNP may be regulated by the gonadal function including the estrogens/androgens circulating ratio; (v) The activity of cardiac endocrine function is depressed by the presence of one or more circulating or tissue bioactive substances (e.g., leptin); (vi) An efficacious (standard) pharmacological treatment of obese individuals or patients with hypertension and/or type-2 diabetes mellitus tends to decrease significantly the production/secretion of BNP and NT-proBNP; (vii) The inverse correlation between BMI and BNP values in patients with heart failure (e.g., with hemodynamic impairment, activation of counteracting neurohormone, increased circulating levels of cytokines) may be partially explained by the effect of the malnutrition-inflammation complex syndrome.
Effects of natriuretic peptide system in obesity, metabolic syndrome and type 2 diabetes (Moro, 2016 and citations therein) [49] (i) Natriuretic peptides promotes fat mobilization and utilization as well as regulate energy metabolism and expenditure; (ii) In obesity, metabolic syndrome and type 2 diabetes there is a state of natriuretic peptide system deficiency (natriuretic handicap); (iii) Contributing factors to the natriuretic handicap are the elevated levels of glucose, insulin, fatty acids, proinflammatory cytokines; (iv) The natriuretic handicap occurs despite a higher left-ventricular mass and end-diastolic pressure in people with obesity compared to lean individuals; (v) Human and mouse adipose tissue express NPR-C; high fat feeding up-regulates NPR-C and reduces NPRA mRNA  [12] (i) Adipokines (e.g., leptin) can differently regulate the production/secretion of ANP and BNP by cardiomyocytes throughout multiple metabolic pathways; these pathways are activated or inhibited depending on different pathophysiological conditions; (ii) A cross talk exits between the cardiac endocrine function and adipose tissue (e.g., ANP and BNP regulate fat tissue function and growth as well as exerting potent lipolytic effects in human fat cells, ANP increases production of adiponectin); (iii) ANP gene expression is increased in a dosedependent manner by female sex steroids; (iv) Estradiol and progesterone are required for normal ANP gene expression by rat cardiomyocytes; (v) In postmenopausal women, a stimulatory action on the production/secretion of ANP and BNP is induced by hormone replacement therapy with female steroid hormones; (vi) There is debate on the effects of androgens on the production/secretion of ANP and BNP (e.g., in atrial cultured myocytes of newborn rats, the synthesis and secretion of both ventricular and atrial ANP is stimulated by testosterone; whereas in castrated male rats, the concentration of ANP in plasma and ANP atrial stores is increased; testosterone replacement reduces ANP levels in plasma, but not in ANP atrial stores; blocking the androgen receptor and, to a lesser extent, suppressing androgen in men with prostate cancer, increases NH2-terminal-proBNP levels in plasma; other studies [e.g., the Dallas Heart Study] in humans suggest that estrogens may stimulate whereas androgens inhibit the production/secretion of ANP and BNP).
Box 3 On the mechanism that regulates the cardiac-specific origin of sPLA2  [56][57] . These specific immune cells scavenge and consequently reduce the levels of free CCL7 below the threshold necessary for induction of sPLA2 by hepatocytes. This would not occur in the heart of Mmp2 -/mice because of lack of significant immune cell homing [8] (as depicted in the enclosed Figure to the right, bottom). Published and unpublished observations by the authors support this notion including that: (i) CCL7 induces the transcription of classical sPLA2 isoenzymes as well as sPLA2 activity in Hepa-1c1c7 cells. (ii) CCR-expressing CD45 + immune cells are found homing to the liver (but not the heart) of Mmp2 -/mice, compared to Mmp2 +/+ mice [8] . -

Box 2 On the molecular constituents of cardiac sPLA2
Since the first identification of its activity [7] , the molecular identity of the MMP-2-regulated cardiac specific secreted PLA2 has remained elusive [7][8] . However, application of an unbiased quantitative proteomics approach could speed up the identification and enable quantitation of cardiac sPLA2 in biological samples. Proposed pillars of a quantitative proteomics approach to identify cardiac sPLA2 from Mmp2 -/mice include: (i) Use stable isotope labeling by amino acids (SILAC) [52][53][54][55] labeling integrated into the isolation strategy. (ii) BLAST-screen the identified cardiac sPLA2 against all known mouse and human PLA2s. (iii) Pursue a targeted "expression cloning strategy" where the cDNA of the identified enzyme is first expressed in, e.g., HEK-293, cells and next CCL7-induction of PLA2 activity is measuredthe hypothesized pathway by which CCL7 elicits sPLA2 release [8] is depicted in Box 3.
inflammatory cytokine which is normally cleaved and inactivated by MMP-2) serves as stimulus for cardiacspecific release of sPLA2 activity [8] . This notion is supported by (a) ex vivo assays data [7] , showing that CCL7 stimulates sPLA2 release from cardiac, but not hepatic tissue and (b) the normalization of the cardiohepatic lipid metabolic phenotype of MMP-2-deficient mice injected with neutralizing CCL7 antibodies, but not with isotype-matched non-immune IgG [8] . However, since CCL7 receptors are expressed on immune cells, cardiomyocytes and hepatocytes [56][57] , it is paradoxical that the liver of Mmp2 -/mice does not exhibit elevated sPLA2 activity, whereas the heart of Mmp2 -/mice does, compared to wild-type mice. In Box 3 we propose a mechanism that may clarify what makes cardiac sPLA2 "cardiac" in origin.
Influence of the heart-centric MMP-2/CCL7/ sPLA2 axis on lipid metabolism A still-open question is whether MMP-2-mediated proteolysis of cytokines, such as CCL7, perturbs lipid metabolism via CCL7-receptor signaling pathways? To answer this question, Box 4 describes two pathways by which the heart influences hepatic lipid metabolism and inflammation.
Future studies will provide precision to the first pathway described in Box 4, including the molecular identity (amino acid sequence) of the enzyme isoforms responsible for cardiac sPLA2 activity (Box 2) and deciphering the mechanism that regulates the cardiac-specific origin of sPLA2 in MMP-2 deficiency (Box 3). Box 4 On the MMP-2 regulated CCL7-mediated modulation of lipid homeostasis MMP-2 cleavage of CCL7 could influence lipid homeostasis through at least two pathways: Postulated pathway 1: Modulation of prostaglandin production by the CCL7 / cardiac sPLA2 axis [7][8] . (i) Cardiac sPLA2 activity, whose release is stimulated by CCL7, reaches non-cardiac organs, such as the liver, through the circulation; (ii) Cardiac sPLA2 hydrolyses membrane phospholipids in the liver releasing the fatty acid esterified at carbon-2 (typically, arachidonic acid, C20:4) and leaving behind a lysophospholipid (e.g., lyso-PC); (iii) Arachidonic acid is next converted into prostaglandin E2 by cyclooxygenases (as depicted in Box 3). Postulated pathway 2: Modulation of cholesterol metabolism by the CCL7 / tumor necrosis factor (TNF)-a /adenosine monophosphate-activated protein kinase (AMPK) axis. As illustrated, CCL7 can influence cholesterol homeostasis in target organs, including the heart and liver. Several key signaling events are: (i) CCL7 binds and activates CCR2 [56] . (ii) A classical G-proteincoupled receptor-like calcium-dependent signaling cascade is elicited and pro-inflammatory genes, including TNF-α, are activated. (iii) Transcriptional induction of protein phosphatase 2C (PP2C) expression is stimulated; (iv) PPC2 reduces phosphorylation of active (phosphorylated) AMPK [58] ; however, reduced AMPK phosphorylation can also be provoked by an interaction between AMPK and phosphoinositide 3-kinase enhancer A-a ubiquitously expressed GTPase and proto-oncogeneas recently reported [59] .