Interactions Networks for Primary Heart Sarcomas

Simple Summary Cardiac cancer represents a rare, largely understudied disease. The aim of this study was to elucidate the underlying pathophysiology of cardiac sarcomas by employing specific network-based methods. Focused interactomes comprised of heart- and tumor-associated gene/proteins were constructed. These networks allowed us to unfold the main pathways leading from heart sarcomas to cardiac diseases. The findings of this study could be utilized in the clinical setting for diagnostic and therapy decision-making. Abstract Personalized medicine incorporates genetic information into medical practice so as to optimize the management of chronic diseases. In rare diseases, such as heart cancer (incidence 0.0017–0.33%), this may be elusive. Ninety-five percent of the cases are due to secondary involvementwith the neoplasm originating in the lungs, breasts, kidney, blood, or skin. The clinical manifestations of heart tumors (benign or malignant) include heart failure, hypertension, and cardiac arrhythmias of varying severity, frequently resulting in blood vessel emboli, including strokes. This study aims to explain the pathophysiology and contribute to a P4 medicine model for use by cardiologists, pathologists, and oncologists. We created six gene/protein heart-related and tumor-related targets high-confidence interactomes, which unfold the main pathways that may lead to cardiac diseases (heart failure, hypertension, coronary artery disease, arrhythmias), i.e., the sympathetic nervous system, the renin-angiotensin-aldosterone axis and the endothelin pathway, and excludes others, such as the K oxidase or cytochrome P450 pathways. We concluded that heart cancer patients could be affected by beta-adrenergic blockers, ACE inhibitors, QT-prolonging antiarrhythmic drugs, antibiotics, and antipsychotics. Interactomes may elucidate unknown pathways, adding to patient/survivor wellness during/after chemo- and/or radio-therapy.


Introduction
Since 1999, when Wulff disputed Newton's mechanistic point of view on 'disease', adopting the Aristotelian one (that an organism is a complex of qualities rather than quantities), complexity became an ongoing research subject in medicine and epidemiology [1]. Thus, the term 'disease' seems to be redefined by summarizing data from various directions (lifestyle, inherited predispositions, medical history, sensory data, imaging, all -omics) [2]. Noteworthy strides in the field have come from oncology, as well as from cardiology, where complexity and heterogeneity are recognized as dominant features. Most importantly,

Results
We constructed six different interaction networks-one for each type of primary heart sarcomas: (i) angiosarcoma, (ii) undifferentiated pleomorphic sarcoma, (iii) Rhabdomyosarcoma (iv) Leiomyosarcoma, (v) Myxofibrosarcoma, and (vi) Synovial sarcoma. The constructed interactomes (henceforth called CS1-6), which include nodes of gene/gene products of known and/or predicted interactions, are described in Table 1 and are illustrated in Figures 1-6. A confidence level > 0.7 was adopted. The names of all nodes are shown in the second column of Table 1. The NCBI RefSeq accession numbers [12] are added in a third column, while, in a fourth column, the specific interactome in which each gene is involved, is listed.

Discussion
Heart failure, coronary artery disease, and hypertension epidemics are major public health problems of high prevalence, whose pathophysiology has been insufficiently examined, especially when implicated in other diseases [13]. Heart cancer, on the other hand, is a rare neoplastic entity, that diagnosticians cannot decipher or evenly aptly treat, due to reasonable lack of familiarity [8]. The relevant research demands considerable costs and its effectiveness is time-and population size-dependent. Hence, in silico analysis as gene/gene products interactions networking (interactome construction) is a sine qua non cost/time-effective way to fill in gaps in knowledge in the little known and highly complex field of heart cancer and its cardiological consequences [4]. The latter include heart failure, hypertension, and cardiac arrhythmias, leading to peripheral emboli or even strokes.
In a previous work, we had constructed the heart cancer interactions network (HFCC1) [14,15]. This work is an expansion of that preliminary work (HFCC1), as it is more focused (on sarcomas) and more specialized. It involves six different interaction

Discussion
Heart failure, coronary artery disease, and hypertension epidemics are major public health problems of high prevalence, whose pathophysiology has been insufficiently examined, especially when implicated in other diseases [13]. Heart cancer, on the other hand, is a rare neoplastic entity, that diagnosticians cannot decipher or evenly aptly treat, due to reasonable lack of familiarity [8]. The relevant research demands considerable costs and its effectiveness is time-and population size-dependent. Hence, in silico analysis as gene/gene products interactions networking (interactome construction) is a sine qua non cost/time-effective way to fill in gaps in knowledge in the little known and highly complex field of heart cancer and its cardiological consequences [4]. The latter include heart failure, hypertension, and cardiac arrhythmias, leading to peripheral emboli or even strokes.
In a previous work, we had constructed the heart cancer interactions network (HFCC1) [14,15]. This work is an expansion of that preliminary work (HFCC1), as it is more focused (on sarcomas) and more specialized. It involves six different interaction networks, one for each heart sarcoma (CS 1-6) ( Table 1).
Seventeen novel genes were detected:  [46] 13. high mobility group AT-hook 2 which is a transcription factor related to malignancy and poor prognosis 14. ISG15 ubiquitin like modifier-a small ribosomal subunit is a multi-modal unit implicated to immune response and more interestingly to cancer stem cells in a tumor [47] 15. nibrin, a protein involved in DNA damage repair [48], whilst its variants were shown to be associated with breast cancer [49] 16. nerve growth factor receptor-a neurotrophic factor involved in the regulation and survival of certain neurons and pancreatic beta cells [50] 17. platelet and endothelial cell adhesion molecule 1-a protein that removes aged neutrophils from the body, involving in leucocyte transmigration and angiogenesis [51].
Four non-mediating genes were recognized, namely: GRAM domain containing 1B (involved in cholesterol metabolism and oncogenesis), KH domain containing 3 like (involved in oncogenesis), subcortical maternal complex member (implicated in maternal effect of unknown exact function), unc-51 like kinase 4 (associated to bipolar disorder and schizophrenia).
The created interactomes elucidate the main pathways leading to cardiac diseases (heart failure, hypertension, coronary artery disease, arrhythmias), such as the reninangiotensin-aldosterone system (RAAS) and endothelin system, and excludes others, such as K oxidase or cytochrome P450 pathways in these patients. [52][53][54][55]: Based on the Mestroni et al. pharmacogenetics study, we noted that only EDN1 and ACE are included in the created interactomes. Although our study is preliminary, this finding is important for personalized medicine, but still needs to be validated in clinical settings in the future.
Angiosarcomas represent 40% of the cardiac sarcomas, are usually found in the right atrioventricular groove (Figure 7) [56], and often expand to the right atrial wall and pericardium. According to the 2015 WHO classification of tumors of the heart and pericardium, angiosarcomas ICD-O code is 9120/3 [56]. The mean prevalent age of heart sarcomas is 41 years, with angiosarcomas accounting for 37% of the total cardiac sarcomas.
in the created interactomes. Although our study is preliminary, this finding is important for personalized medicine, but still needs to be validated in clinical settings in the future.
Angiosarcomas represent 40% of the cardiac sarcomas, are usually found in the right atrioventricular groove (Figure 7) [56], and often expand to the right atrial wall and pericardium. According to the 2015 WHO classification of tumors of the heart and pericardium, angiosarcomas ICD-O code is 9120/3 [56]. The mean prevalent age of heart sarcomas is 41 years, with angiosarcomas accounting for 37% of the total cardiac sarcomas. The constructed interactome (henceforth called CS1) illustrated in Figure 1 involves 61 nodes. Its major hubs are IL-6, FGF2, TP53, KRAS, CCND1. The following genes are expressed exclusively in angiosarcomas and not in the rest sarcoma types: CDKN2A, CDKN2B, EP300, KIT, NRAS, POT1, NBN, KDR, KMT2D, PPLG1. Our study confirms the findings of literature references [57]. Notably, KRAS is expressed in angiosarcoma and rhabdomyosarcoma as well.
We revealed novel not previously reported interactions in CS 1: platelet and endothelial cell adhesion molecule 1 (PECAM1) and nibrin (NBN).
Moreover, we identified one case of this extremely rara avis cancer entity (angiosarcoma) in the Onassis Cardiac Surgery Center in Athens, which is presented for the first time ( Figure 7): a Greek woman < 41 years of age within the established in literature age range. Of note, our in silico study is not based on the individual data of this case. The constructed interactome (henceforth called CS1) illustrated in Figure 1 involves 61 nodes. Its major hubs are IL-6, FGF2, TP53, KRAS, CCND1. The following genes are expressed exclusively in angiosarcomas and not in the rest sarcoma types: CDKN2A, CDKN2B, EP300, KIT, NRAS, POT1, NBN, KDR, KMT2D, PPLG1. Our study confirms the findings of literature references [57]. Notably, KRAS is expressed in angiosarcoma and rhabdomyosarcoma as well.
We revealed novel not previously reported interactions in CS 1: platelet and endothelial cell adhesion molecule 1 (PECAM1) and nibrin (NBN).
Moreover, we identified one case of this extremely rara avis cancer entity (angiosarcoma) in the Onassis Cardiac Surgery Center in Athens, which is presented for the first time ( Figure 7): a Greek woman < 41 years of age within the established in literature age range. Of note, our in silico study is not based on the individual data of this case.
The undifferentiated pleomorphic sarcoma is classified with the code number 8830/3 in the WHO classification of tumors of the heart and pericardium of 2015 [56].
Rhabdomyosarcoma classification number in the ICD-O Classification of Diseases for Oncology of 2015 is 8900/3 [56]. It is a tumor type routinely found in infants even fetuses, thus, assumed to be a congenital hamartoma. No sexual dimorphism has been reported, while it is usually localized in the ventricular myocardium or every so often project into the heart cavity. Homogenous echogenicity characterizes imaging of these tumors, whilst they incite arrhythmias. The rhabdomyosarcoma (hence called CS3) interactions network numbers 46 nodes, where IL-6 and FGF2 are its most connected nodes. CS3 is described in Figure 3. KRAS is involved both in angiosarcomas and rhabdomyosarcomas. It straightforward connects to PDGFR13, PDGFRA, FGF2, FGFR1, CCND1, and IL-6.
Leiomyosarcoma ICD-O code number in the ICD-O Classification of Diseases for Oncology of 2015 is 8890/3 [56]. This type of sarcoma is typically found in the left atrium, and has been characterized by specific tissue differentiation. The relevant interactome subsequently referred to as CS4 is shown in Figure 4 and contains only 46 nodes. It is the smallest network constructed herein, whilst its major hubs are IL-6, FGF2, and HRAS. The latter is expressed exclusively in this type of heart sarcoma, and, thus, may be assumed as typical of the entity. FGF2 interacts directly with IL-6 that interacts in its turn with HRAS. The latter connects directly to EDN1 or CCND1, FGFR1, MMP2, and VWF.
Myxofibrosarcoma is coded with the number 8811/3 in the ICD-O Classification of Diseases for Oncology of 2015 [56]. It was previously called malignant fibrous histiocytoma with at least one-quarter of the myxoid areas representing most of the left atrial sarcomas. The network created is henceforth called CS5 ( Figure 5) and involves 62 nodes, while it is the most crowded network presented herein. Its major hubs are IL-6, FGF2, and MMP2.
This network is the only one that involves RHOB-a cancer cell apoptosis modulator and, currently a research target as a cancer therapeutic. As apoptotic factor, it decreases while tumors proliferate, whilst it has been identified in head and neck, lung, and brain cancers and in adenocarcinomas, as a poor prognosis indicator.
LGALS3 interacts only with MMP1, MMP2. CYB5A is linked to SCD, except DECR1. ISG15 and IF16 are directly linked together but through PCNA and then through CCND1 they connect indirectly to the common to all sarcomas' major hubs FGF2 and IL-6. SMS connects to ASS1 which connects to ADSS and then to NKX2-5, which through GATA4 affects EDN1.
In keeping with the ICD-O Classification of Diseases for Oncology of 2015, the synovial sarcoma's code number is 9040/3 [56]. The interactions network we have built (henceforward called CS6) includes 47 nodes and is shown in Figure 6. The high degree connections of this specific interactome are IL-6 and FGF2. Notably, HRAS, SS18, SSX2, MVP2 are uniquely expressed in synovial heart sarcoma. Of those, SSX2 and SS18 have been identified in histology-type investigations in the literature, whilst MVP2 function is little known, while HRAS is implicated in various types of cancers (i.e., salivary duct carcinoma [58], epithelial myoepithelial carcinoma [59], etc. In sum, we observed that all cardiac sarcomas share • forty-four common nodes • two common hubs: IL-6 and FGF2, that may be assumed as typical of the overall entity (heart sarcomas) All six interactions networks are presented in 3 basic color-specified units/clusters. Nodes in a given cluster are most closely connected to those nodes of the corresponding cluster, as compared to the other two clusters. Nodes in the same cluster module can be associated with similar/common biological functions. This information could be expanded to more complicated enriched clusters that represent cardiac sarcoma-related biological processes, in the future.
Finally, the wild-type gene inclusion is a limitation in this investigation.

Materials and Methods
The protocol followed is the established one in molecular networking [60,61] and is described in Figure 8.

Conclusions
Cardiac sarcoma patients might profit from administration of beta-adrenergic blockers, ACE inhibitors, QT-prolonging antiarrhythmic drugs, antibiotics, and antipsychotics. Interactomes may elucidate unknown cardiac implications of cardiac malignancies and contribute to patient/survivors' wellness during and after chemo and/or radio treatment.  The interactions among gene/gene products were studied through STRING v10 [62]. A relative high confidence interaction score of 0.7-0.97 was opted so as to extract relevant information from multiple and diverse sources, avoiding at the same time the inclusion of rather erroneous interactions (e.g., false positives). The average degree of connectivity (i.e., mean number of connections of a given node to its immediate neighbors) and kmeans clustering in the constructed networks were computed with the usage of algorithms implemented in the STRING database. The nodes in the network were partitioned into three clusters, based on the shortest path (distance) between two given nodes. Intracluster edges are represented by solid lines, whereas inter-cluster edges are denoted by dashed lines.
Genes not previously published to be associated to the specific heart sarcoma type, thus, not been included in the initial set of molecules in String platform so as to be processed are described as 'novel' or 'novelties' in this text.
All six interaction networks are presented in three basic color-specified units/clusters. Nodes in a given cluster are most closely connected to those nodes of the corresponding cluster, as compared to the other two clusters. Nodes in the same cluster module can be associated with similar/common biological functions.

Conclusions
Cardiac sarcoma patients might profit from administration of beta-adrenergic blockers, ACE inhibitors, QT-prolonging antiarrhythmic drugs, antibiotics, and antipsychotics. Interactomes may elucidate unknown cardiac implications of cardiac malignancies and contribute to patient/survivors' wellness during and after chemo and/or radio treatment.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.