ANTs-induced cardiotoxicity has been extensively studied for over 50 years, yet the mechanisms remain unclear. A prominent pathological feature of DOX cardiotoxicity is cardiac fibrosis, leading to congestive dilated cardiomyopathy and congestive heart failure [16]. The detailed mechanisms of DOX cardiotoxicity are being refined and explored. Current studies suggest that potential mechanisms of DOX cardiotoxicity include DNA damage, ROS accumulation, iron toxicity, lipid peroxidation, apoptosis, energy deficiency, mitochondrial disruption, autophagy, necrosis, mutagenesis, and calcium disorders [17]. This study had two objectives: (1) to discover potential targets and mechanisms of DOX cardiotoxicity and (2) to find the targets of saffron against DOX cardiotoxicity based on the discovery of DOX cardiotoxicity targets.
4.1 The targets and mechanisms of DOX cardiotoxicity
One of the objectives of this study was to discover the targets and mechanisms of DOX cardiotoxicity. An unexpected finding was the significant enrichment of ribosomal proteins (MRPS7, RPS5, MRPL19, MRPS2, MRPL21, MRPL9, MRPL23, and MRPL16) in the DOX group. Therefore, we hypothesized that ribosomes are involved in the pathogenesis of DOX cardiotoxicity. The ribosomal pathway has not been widely found in previous studies on the mechanisms of DOX cardiotoxicity. In order to confirm whether this discovery is an accidental phenomenon and has clinical application value, we conducted further research and verification. In this study, we performed complementary transcriptomic analyses to support the results derived from the proteomics data. We found that MRPL4, MRPL19, MRPS9, RPL10, RPL24, RPS15, RPL30, and RPS24 were significantly decreased in the DOX group compared to the CONT group. We performed a comparative analysis using multiple methods and found that the ribosomal pathway was significantly enriched in both transcriptomic and proteomics results. Although the combination of proteomics and transcriptomic is synergistic for identifying disease-specific biomarkers, numerous studies have reported inconsistent results between proteomics and transcriptomic data [18, 19]. In our research, there are also some differences between transcriptomic and proteomics. For example, amino acid tRNA biosynthesis, complement and coagulation cascades, the PPAR signal transduction pathway, the mTOR signal transduction pathway, unsaturated fatty acid biosynthesis, and other related pathways are enriched in proteomics but not in transcriptomics. There are three reasons for the differences in results: variability between experimental samples, differences in assay methods, and differences in experimental conditions and operators. It is worth stating that proteins are closer to the phenotype than are transcripts [20]; thus, we used the results of proteomics as a basis and the transcriptomic results only as a reference to assist in validation.
Furthermore, the present study was not limited to the validation of transcriptomic studies on proteomics but rather was also combined with a ribosomal ubiquitination study on DOX cardiotoxicity [15] and more literature to confirm our speculations. This study found that DOX-induced DNA damage resulted in extensive ubiquitination of ribosomal proteins (RPs) [15], of which RPS10, RPS13, RPS19, RPS7, RPL24, RPL18, RPL18A, RPL3, and RPL30 are consistent with the RPs identified in our study. It is encouraging that several studies have confirmed that ribosomes may be involved in DOX cardiotoxicity [21, 22]; therefore, ribosomes could be a direction for future research on DOX cardiotoxicity. So, how are ribosomes involved in DOX cardiotoxicity? Next, we tried to explore the possible mechanisms through a literature review and partial validation experiments.
4.2 Hypothesis on the mechanisms of ribosomal involvement in DOX cardiotoxicity
A ribosome is an organelle commonly found in cells, mainly composed of rRNA and proteins, and the process of translating mRNA into proteins in the ‘Central Dogma’ occurs in ribosomes [23]. Therefore, the ribosome is also known as the intracellular protein synthesis machine. The number of ribosomes in the heart increases as the heart grows since approximately 80% of its RNA is ribosomal [24]. In order to pump blood continuously throughout the body, the heart requires high energy production through many mitochondria. Mitochondrial ribosomes are critical for energy production. Ribosomes are indispensable to the maintenance of cardiac function. Ribosome biogenesis plays a crucial role in cell growth by enhancing protein synthesis [25]. Moreover, ribosome biogenesis is a tightly coordinated and highly dynamic process in which ribosomal RNA (rRNA) is synthesized, modified, and assembled with RPs to form mature ribosomes [26]. However, if ribosome biosynthesis is perturbed, it will lead to nucleolar stress response [27]. On the one hand, free RPs and 5S rRNA are released from the nucleosome to the nucleoplasm and activate the P53 pathway through the MDM2–MDM4–p53 axis, RPS–MDM2 interactions, binding to P53 mRNA via RPL26, and enhancing P53 translation (Fig. 10). Currently, an increasing number of ribosomal proteins have been shown to regulate the MDM2–MDM4–p53 axis, which includes RPL5, RPL6, RPL11, RPL22, RPL23, RPL26, RPL37, RPS3, RPS7, RPS14, RPS15, RPS19, RPS20, RPS25, RPS26, RPS27, RPS27A, and RPS27 [26].
On the other hand, due to impaired pre-rRNA synthesis or processing, ribosomal proteins are rapidly degraded when they fail to assemble into functional ribosomal subunits [28]. It appears that there is a quality control mechanism to ensure that excess free ribosomal proteins do not accumulate, thereby preventing their collision with translationally stalled ribosomes or inappropriate interactions with other cellular components [29, 30]. The ubiquitination of ribosomal proteins is crucial in the quality control of ribosome arrest induction [31]. Notably, we previously raised that when ribosome biogenesis disorders cause ribosomal stress, it leads to a significant transfer of ribosomes from the nucleosome to the nucleoplasm, which activates the P53 pathway. Conversely, ribosomal stress leads to ubiquitination, degrading free ribosomes to prevent free ribosomal protein aggregation.
So, is there a conflict between ubiquitination and activating the P53 pathway? Several studies have answered this question, with the findings of extensive ubiquitination of RPs at 2 and 6 h after the induction of DNA damage [15]. Moreover, it has been demonstrated that the ubiquitination of RPs does not change significantly at earlier time points after damage [32]. Thus, the ubiquitination of ribosomal proteins may be part of the response to DNA damage in the mid-to-late stages.
Furthermore, it is becoming increasingly clear that fully functional ribosomes are necessary for cellular redox homeostasis. Ribosomal proteins can act as sensors for oxidative stress to mediate a decrease in translation [33]. Ribosomal RNA is a target of oxidative nucleobase damage, and increased levels of oxidative stress can interfere with different sub-steps of the ribosome assembly and translation elongation cycle [34, 35] and reduce the fidelity of protein translation [36]. The RPL10-R98S (uL16-R98S) mutation is associated with elevated ROS-mediated oxidative stress and DNA damage [34]. Moreover, damaged or mutated ribosomal proteins can mediate oxidative stress. It has been demonstrated that increased levels of reactive oxygen species (ROS) have been detected in RPL5- and RPS19-deficient Diamond–Blackfan anemia (DBA) mouse cell models [37]. Thus, the disruption of ribosome assembly increases ROS production, thereby leading to an oxidative and translational defective cellular “snowball” effect that may further enhance the mutagenic phenotype [34].
In summary, the following hypothesis regarding ribosomal involvement in the mechanism of DOX cardiotoxicity was proposed: DOX, a topoisomerase II inhibitor, inhibits pre-RNA synthesis during ribosome biology in cardiomyocytes, leading to ribosomal stress. Large amounts of ribosomal proteins and rRNA are transferred from the nucleolus to the nucleoplasm, whereby activating the P53 pathway through the MDM2–MDM4–P53 axis, RPS–MDM2 interaction, and RPL26 binding to p53 mRNA and enhancing P53 translation. P53 activates downstream pathways, leading to cardiomyocyte apoptosis, autophagy, cell cycle arrest, and cardiomyocyte damage. In addition to late DNA damage caused by DOX, ribosomal stress triggers ribosome-related quality control (RQC) and activates ubiquitination, and the ubiquitination process activates oxidative stress. In addition, DOX causes oxidative damage to ribosomal proteins and rRNA through oxidative stress triggered by mitochondrial damage, generating excess ROS, which leads to further disruption of ribosome assembly and function, forming a vicious cycle of ribosome-mediated cardiotoxicity.
Finally, we validated, using ELISA, several critical targets in the hypothesized mechanism of ribosomal involvement in DOX cardiotoxicity (Supplementary Fig. 4). Ribosomal proteins RPL11, RPL10, RPs15, RPL5, RPL26, TP53, and MDM 4 were significantly increased in the DOX group compared to the CONT group. This result is consistent with part of our proposed hypothesis that DNA damage followed by ribosomal stress causes many ribosomal proteins to be transferred from the nucleolus to the nucleoplasm, activating the P53 pathway and causing cardiac injury. However, it should be noted that the ELISA validation results regarding ribosomal proteins are not consistent with the proteomics results. Possible reasons for the analysis include the following: the protein-profiling technique has the problem of inaccurate quantification; therefore, a false positive caused by the assay method cannot be excluded. Furthermore, the ubiquitination of ribosomes is a dynamic and complex process related to various factors, such as the time of detection and subcellular localization of the protein. In addition, mass spectrometry and ELISA detect proteins in-stock quantities, not transiently expressed proteins. Therefore, how ribosomes are involved in DOX cardiotoxicity needs to be investigated in the future.
4.3 The mechanisms of action of saffron extract against DOX-induced cardiotoxicity
Recently, various studies have demonstrated the medicinal properties of saffron, such as antioxidant [38], antitumor [39, 40], antidepressant, anxiolytic, hypnotic [41–43], anti-Alzheimer’s disease [44], gene protection [45, 46], hepatoprotective [47], neuroprotective [48, 49], antiplatelet [50], lipid modulator, anti-atherosclerotic [51, 52], anti-inflammatory [53], and improvement of insulin resistance [54]. Most importantly, saffron and its ingredients reduce the toxicity of natural and chemical agents [55] and have good cardioprotective effects [56, 57]. It has been reported that the newly marketed exclusive and innovative herbal product Saffron Total Glucoside Tablets® has been used to prevent myocardial injury due to DOX. Clinical and laboratory-based studies have shown that saffron has no significant toxicity within therapeutic doses [58]. In short, saffron is safe and effective against DOX cardiotoxicity.
So, what is the active ingredient of saffron against Adriamycin cardiotoxicity? The main components of saffron include crocin, crocetin, picrocrocin and safranal. We performed an HPLC analysis of saffron extract and found that the content of crocin Ⅰ and crocin Ⅱ in the saffron extract was greater than 90% (Supplementary Fig. 5). Furthermore, in combination with the results of our previous network pharmacology study, some components of saffron were found to be closely associated with DOX cardiotoxicity, including quercetin, kaempferol, isorhamnetin, isorhamnetin 3-robinobioside, kaempferol, crocin Ⅰ, isorhamnetin 3,4'-diglucoside, crocin Ⅲ, crocin Ⅱ, isorhamnetin-3-O-glucoside, etc. The combined analysis speculates that crocin Ⅰ and crocin Ⅱ may be the active components of saffron against DOX cardiotoxicity.
So, how does saffron exert its anti-DOX cardiotoxic effect? Although several studies have demonstrated that saffron effectively alleviates DOX cardiotoxicity, the mechanisms thereof remain unclear. Only a few studies have shown the mechanisms of saffron anti-DOX cardiotoxicity [7]. Therefore, the second main objective of this study was to discover the key targets and pathways of saffron anti-DOX cardiotoxicity through proteomics.
Through proteomics analysis, a total of 103 critical targets were screened, of which nine proteins were identified as potential targets of saffron for DOX cardiotoxicity, namely XIRP2, EPHX1, MAPKAPK2, SORBS2, CD81, FLOT2, FLOT1, CD59, and LAMA5. Moreover, it is important to note that, as mentioned in the previous paragraph, we found that the 103 critical targets of DOX cardiotoxicity in the top 10 hub proteins were MRPS9, PTCD3, MRPL4, MRPS30, MRPS27, RPS15, RPL10, HSPA5, CTSD, and DCN, while the nine targets that were finally identified as saffron anti-DOX cardiotoxicity were inconsistent with them. What led to this result? According to the PPI network map of 103 key targets of DOX cardiotoxicity, we found that the identified targets of saffron against DOX cardiotoxicity were free in the PPI network, while the DOX cardiotoxicity targets were screened in the top 10 based on degree values. It is well known that in PPI analysis, degree values indicate the degree of connectivity of protein interactions. Therefore, the two results are inconsistent. Furthermore, why were the top 10 hub proteins screened by the DOX group compared to the CONT group not in the intersecting target dataset in the SE, DOX, and CONT groups? This is because saffron acts on these targets and regulates protein expression; therefore, it does not show the difference. To confirm this idea, we observed whether saffron modulated these targets (CTSD, DCN, HSPA5). The results showed significant differences in these targets in the SE group compared to the DOX group.
Nine proteins were identified as potential targets of saffron for DOX cardiotoxicity. Ten proteins were key targets for DOX cardiotoxicity. Based on the literature, we have selected some of these targets for ELISA validation and will focus on targets closely linked to cardiovascular disease, namely XIRP2, EPHX1, SORBS2, CD81, FLOT2, FLOT1, CD59, DCN, CTSD and HSPA5.
Xin actin binding repeat containing 2 (XIRP2) is a tissue-specific membrane-associated protein (essential for normal cardiac conduction) that regulates the Hippo-YAP signaling pathway and controls cardiac development [59]. Xirp2-null hearts exhibit prolonged PR and QT intervals, slow conduction velocity, AV block, and abnormalities of the proximal ventricular conduction system [60]. One study found reduced expression of Xirp2 in the hearts of dilated cardiomyopathy (DCM) patients [59], but another study showed significantly increased expression of XIRP 1 and XIRP 2 in a mouse model of dilated cardiomyopathy injected with tamoxifen for two weeks [61]. It is speculated that this difference may be due to differences in the disease stage, which is thought to be relatively early in the pathological progression of DCM [59]. This matched our findings, which showed that XIRP2 expression was increased after two weeks of DOX injection. The role of these proteins in the progression of cardiac disease warrants further investigation, and XIRP2 might be a novel therapeutic target for cardiovascular disease and a key target of saffron against DOX cardiotoxicity.
Microsomal epoxide hydrolase (EPHX1) is an enzyme involved in protecting cells against oxidative stress [62]. EPHX1 is widely accepted as playing a dual role in PAHs and aromatic amines detoxification and activation [63]. EPHX1 expression was reduced in a study of RNA-Seq datasets from transplanted hearts of patients with heart failure [64]. In contrast, some studies suggested that EPHX1 is involved in diseases such as atherosclerosis, sepsis, stroke, and myocardial infarction, and that inhibition of EPHX1 may be beneficial in the acute treatment of arrhythmias and sudden cardiac death [65]. In our study, we found that EPHX1 was significantly elevated in the DOX group relative to the CONT group. While EPHX1 showed a downward trend in the SE group relative to the DOX group, but there was no statistical difference. It is suggested that EPHX1 may also be associated with the DOX cardiotoxicity, but further in-depth studies are needed.
Sorbin and SH3 domain containing 2 (SORBS2), which is subcellular and localized in epithelial and cardiac myocytes, acts as an adapter protein that may be associated with the precordial disease. SORBS2 plays a role in coordinating multiple signaling pathways converging on actin and microtubule cytoskeletons [66, 67]. Recent studies have demonstrated the pathogenic effect of elevated SORBS2 levels upon left ventricular noncompaction cardiomyopathy (LVNC), where SORBS2 interacts with β-tubulin and promotes microtubule densification, elevated SORBS2 levels significantly disrupt cardiac mechanical function, and dysregulated SORBS2 accumulation disrupts the spatial organization and perhaps the excitation-contraction coupling (ECC)-a related function of junctophilin 2 [66]. It is believed that SORBS2 has the potential to be of diagnostic value for the early detection or ongoing management of cardiac disease [66]. Our study found that SORBS2 levels were significantly upregulated in the DOX group relative to the CONT group, while SORBS2 was significantly decreased in the SE group relative to the DOX group. SORBS2 may be related to the cardiotoxicity of DOX, and saffron may exert therapeutic effects by regulating SORBS2.
Lipid rafts and lipid raft proteins function in the cell membrane for signal transduction, membrane transportation, and cell adhesion [68]. Transducer signaling is increased if specific lipid raft proteins are expressed, and excessive signals promote tumorigenesis and development. Flotillin 1 (FLOT1) and flotillin 2 (FLOT2) are lipid raft markers [69]. The association between FLOT1 and FLOT2 targets and cardiotoxicity has not been reported. However, it has been demonstrated that FLOT1 and FLOT2 are upregulated in expression and play a regulatory role in various malignancies. They can play a pro-cancer role by activating NF-κB and β-catenin signaling pathways. In this study, FLOT1 and FLOT2 were significantly elevated in the DOX group compared to the CONT group and significantly downregulated in the SE group compared to the DOX group. This finding suggests that FLOT1 and FLOT2 may be related to DOX cardiotoxicity, but the related linkage mechanisms still need future in-depth study.
CD59, an extracellular cell membrane-attached protein, inhibits the formation of the membrane attack complex (MAC), a major effector of complement-mediated tissue damage [70]. Several studies have shown that deletion (mutation) or reduced expression of CD59 increases the risk of cardiovascular or neurological lesions [71–73][68–70]. In animal models, ablation of CD59 promoted accelerated atherosclerosis with occlusive coronary artery disease and premature death, which were attenuated by transgene overexpression of human CD59 in the endothelium [74]. In our study, CD59 was found to be significantly upregulated in the hearts of rats in the DOX group, CD59 was slightly elevated in SE group relative to the DOX group, but not statistically different. It is suggested that CD59 may also be associated with the DOX cardiotoxicity.
Ctsd encodes cathepsin D (CTSD), a lysosomal aspartyl protease that is essential for protein degradation, as well as protein hydrolysis activation of hormones and growth factors, and is secreted mainly after oxidative stress [75, 76]. Myocardial CTSD is downregulated in sudden cardiac death [77]. Circulating cathepsin D levels are associated with HF severity and poorer outcome, and reduced levels of cathepsin D may have detrimental effects with therapeutic potential in HF [78]. Our study found that CTSD was significantly lower in the DOX group than in the CONT group and significantly higher in the SE group than in the DOX group. Moreover, CTSD may be closely associated with DOX cardiotoxicity as well as the therapeutic effects of saffron.
Decorin (DCN) is widely known for its effects on collagen fibril production and the maintenance of tissue integrity. DCN can be released as a secreted proteoglycan under pathological conditions, and serum DCN is a novel biomarker of acute coronary syndrome that contributes to the increased inflammatory response in ischemic heart disease [79]. Our study found that DCN was significantly upregulated in the DOX group compared to the CONT group, while DCN was significantly downregulated in the SE group compared to the DOX group. It is suggested that DCN may also be associated with DOX cardiotoxicity, while saffron may also improve DOX cardiotoxicity by modulating DCN.
Heat shock protein A5 (HSPA5), a member of the heat shock protein family A (Hsp70), is a host cell chaperone protein and a regulator of endoplasmic reticulum (ER) function [80, 81]. This protein maintains protein homeostasis and Ca2+ balance in the ER, is responsible for unfolded protein responses in the ER, and is involved in ER stress-induced apoptosis and autophagy [82, 83]. Currently, HSPA5 is a target for the protection of cardiomyocytes from apoptosis [84]. Because it is an oxidative stress sensor and responder, it can directly protect cells from ER stress and reactive oxygen species (ROS)-induced damage [85]. A proteomics analysis of ANTs-induced cardiotoxicity demonstrated that the influence of ANTs immediately reduced the expression of HSPA5 [82]. Therefore, HSPA5 failed to protect cardiomyocytes, thus leading to chronic damage of cardiac tissue [82]. Our study found that the levels of HSPA5 were significantly downregulated in the DOX group compared to the CONT group. In contrast, HSPA5 was significantly upregulated in the SE group compared to the DOX group. This result further confirmed the relationship between HSPA5 and the cardiotoxicity of DOX. And saffron may also improve the cardiotoxicity of DOX by upregulating HSPA5.
CD81 is a transmembrane protein with multiple biological activities, widely expressed in human tissues, mediating signal transduction and cell activation and growth [86]. It has been demonstrated that CD81 contributes to tumor growth and metastasis, is expressed in most types of cancer, and its expression levels correlate with prognosis in tumor patients [87]. There is no direct evidence linking CD81 to cardiovascular disease or cardiotoxicity. However, several studies have shown exosomes using CD81 as a marker to have cardioprotective effects [88, 89]. It has been demonstrated that exosomes promote vascular endothelial formation and proliferation through activation of Akt (protein kinase B) and ERK (extracellular signal-regulatedkinase), enhancing myocardial repair [88]. Exosomes have also been demonstrated to deliver endogenous protective signals to the myocardium via the pathway of TLR4 and classical cardioprotective HSP pathway [89]. We found that the levels of CD81 were significantly increased in the DOX group and significantly downregulated in the SE group. Thus, CD81 may be associated with DOX cardiotoxicity and saffron modulation regulates CD81 anti-DOX cardiotoxicity.
However, there is the third limitation of this paper: due to the inaccuracy of the quantification of proteomics, the ELISA validation results regarding some essential targets are somewhat different from the proteomics; therefore, more in-depth studies are needed in order to confirm them in the future. In addition, some targets (such as SORBS2 and XIRP2) are undeniably valuable in diagnosing and treating cardiovascular diseases, but some current studies are controversial about these targets; thus, their association mechanisms with DOX cardiotoxicity also need to be further investigated. Third, in the future, a comprehensive interpretation of the complex link between ribosomes and DOX cardiotoxicity and the mechanisms of the anti-DOX cardiotoxic effect of saffron through the ribosomal pathway is needed in conjunction with translationomics. The elucidation of these issues has important implications for the prevention and treatment of ANTs cardiotoxicity.