Drug-induced Cardiotoxicity (DICT) has become a major challenge limiting drug application scenarios, most typically with antineoplastic drugs (Barachini et al., 2023). Among them, anthracyclines were first noted to induce cardiotoxicity manifested as progressive heart failure (Cardinale et al., 2020). While the introduction of more precisely targeted drugs was initially intended to lead to a better prognosis, some unanticipated cardiovascular adverse events have certainly cast a dark shadow. In parallel with the benefits of trastuzumab, which targets epidermal growth factor receptor-2 (HER2), 27% of patients experience cardiac dysfunction, and 16% experience symptomatic heart failure (Slamon et al., 2001). Extensive reviews of breast cancer survivors indicated a greater risk of death from cardiovascular disease, exceeding their initial risk of death from cancer itself, especially in older patients (Chen et al., 2012; Siegel et al., 2012). In addition to this, immune checkpoint inhibitors have been shown to be associated with inflammation-related cardiovascular adverse events, particularly myocarditis (odds ratio 11.21, 95% CI 9.36–13.43) (Salem et al., 2018; Varricchi et al., 2018b). Statistically, up to 27 drugs have been forced off the market in the last few decades due to severe cardiotoxicity (Mamoshina et al., 2021). Hence, elucidating the mechanisms of DICT is of great significance for the broader application of existing agents and the development of new drugs. The plethora of unsolved issues has even spawned the emerging discipline of cardio-oncology (Bellinger et al., 2015).

Various mechanisms concerning DICT have been elucidated, of which mitochondrial dysfunction (Rocca et al., 2022), activation of apoptotic signaling (Christidi and Brunham, 2021), and calcium channel dysfunction (Herrmann, 2020) have been widely recognized. Several reviews have already summarized the mechanisms of DICT, establishing a bridge from microscopic reactions to macroscopic effects and explaining how specific mechanisms of cardiotoxic drugs lead to cardiotoxicity (Herrmann, 2020; Morelli et al., 2022; Abdul-Rahman et al., 2023) (Table 1).

Behind these mechanisms, unexpected drug-protein interactions play a crucial role. Protein homeostasis, abbreviated as proteostasis, is the dynamic equilibrium of protein concentration, localization, and conformation within any living cell. Dependence on molecular chaperones, proteolytic machinery, endoplasmic reticulum (ER), and their regulators enables protein synthesis, folding, and degradation to occur under scrutiny (Hoppe and Cohen, 2020). However, under the stress of changing metabolic, environmental, and pathological conditions, unfolded or unfolded proteins accumulate, thus constantly jeopardizing proteostasis (Hipp et al., 2019). A growing body of evidence indicates that diseases such as Alzheimer’s and diabetes are associated with the accumulation of misfolded proteins (Cozachenco et al., 2023). Among the numerous identified risk factors, pharmaceutical agents, particularly antineoplastic drugs, present a significant challenge to proteostasis. Tumor cells rely on a high degree of proteostasis to cope with their genetic instability and the tumor microenvironment, and a considerable number of antineoplastic drugs target this feature (Liang et al., 2023). However, once off-target effects arise, they can lead to irreversible adverse consequences for relatively vulnerable normal cells. Therefore, resolving the alterations in the conformation and stability of proteins in cardiac cells (cardiomyocytes, endothelial cells, immune cells, etc.) induced by drugs is essential for targeting and rescuing cardiotoxicity and expanding the scope of application of existing drugs (Wang et al., 2020).

Evidence from in vitro experiments demonstrates that various denaturants, such as heat, chemicals, pressure, and force, cause distinct unfolding processes (Lapidus, 2017). Similarly, pharmaceuticals can induce the unfolding process by functioning as chemical denaturants in vivo (Ayaz et al., 2023). In addition to this, endoplasmic reticulum stress, molecular chaperone dysfunction, proteasome dysfunction, and impaired autophagy pathways together contribute to perturbed proteostasis in the heart. By reviewing recent advances in the study of drug-induced cardiotoxicity, we aimed to provide a novel insight into the drug-protein interaction in vivo. Together with the evaluation of in vivo monitoring techniques and targeted interventions, we explored the potential of proteostasis in preventing DICT.

Current computational techniques that rely on sequence similarity have been quite effective in predicting the native structures of proteins (Jeppesen and André, 2023). However, these methods do not adequately tackle the task of identifying folding intermediates. This situation is regrettable since comprehending the folding and unfolding routes of specific proteins is crucial for our understanding of many proteostasis disorders (Sari et al., 2024). The initial research into the mechanism of polypeptide chain folding during protein synthesis was primarily concerned with thermodynamic aspects, such as the calculation of folding free energy. Nobel laureate B. F. Anfinsen postulated in his work on the spontaneous refolding of bovine pancreatic ribonuclease—a protein with a disulfide bond—that denatured proteins could be refolded in vitro in the 1960s (Vila, 2023). Typically, in vitro tests involve treating the protein with substances like urea, guanidine hydrochloride, or heat to unfold, and then removing these substances to allow the protein to regain its original structure (Seelig and Seelig, 2023). In addition to X-rays, nuclear magnetic resonance (NMR) (Zhuravleva and Korzhnev, 2017) and circular dichroism (CD) spectroscopy (Singh et al., 2015), techniques that tend to provide structural information, advanced experimental techniques, such as single-molecule fluorescence resonance energy transfer (FRET), may effectively determine the duration of transition path times and therefore provide insights into the timing of important folding processes (Barth et al., 2019).

As for living organisms, a sophisticated network of molecular processes is engaged to ensure the integrity and functionality of the proteome (Figure 1). Protein biogenesis is achieved through the recognition of nascent polypeptide chains within the ribosomal tunnels and the utilization of molecular chaperones to ensure the accurate folding of newly translated polypeptides (Burke et al., 2022). The endoplasmic reticulum is the site where post-translational modification and protein folding occur. Folding abnormalities of nascent peptides in ER due to the interference of endogenous and exogenous factors will cause the accumulation of unfolded and misfolded proteins, which will ultimately cause ER stress (Oakes and Papa, 2015). Once misfolded proteins are further accumulated, the unfolded protein response (UPR) will be activated to reduce ER burden by enhancing the endoplasmic reticulum-associated protein degradation (ERAD), thus maintaining proteostasis (Senft and Ronai, 2015). Fully processed proteins are then transported to their respective organelles, where they engage in specific enzymatic or structural functions. In the event of protein damage or the necessity for protein recycling, these proteins are degraded through two main processes: the ubiquitin-proteasome system and autophagy. Parts of the misfolded proteins can be refolded by specialized molecular chaperones or, if not, eventually degraded by the proteasome. While the other portion of misfolded protein proteins that are prone to aggregation accumulate to form toxic oligomers and even pathogenic aggregates (Pohl and Dikic, 2019).

Significant progress has been achieved in recent years in the visualization and monitoring of intracellular proteostasis. In essence, these methods can be classified into four principal categories: optical-based techniques, mass spectrometry-based techniques, intracellular nuclear magnetic resonance (NMR) techniques and cryo-electron microscopy techniques (Bai et al., 2023). The observation of oligomers and protein aggregates can be achieved through the utilization of specific fluorescent probes, such as Tetraphenylethene Maleimide (TPE-MI) developed by (Tang et al., 2021). Mass spectrometry-based methodologies offer a distinct advantage in the identification of individual protein species, as they permit the visualization of the protein structure (de Souza and Picotti, 2020). Moreover, intracellular NMR techniques afford a comprehensive insight into the dynamic behavior of a single protein, both prior to and following a misfolding event (Luchinat et al., 2022). The rapid evolution of cryo-electron tomography techniques demonstrates considerable potential for visualizing the structure and properties of protein aggregates within complex and crowded intracellular environments (Watanabe et al., 2020). The aforementioned technical tools afford researchers an intuitive view of how drugs affect proteostasis.

Drugs usually bind to certain proteins in vivo, changing the target protein’s and downstream proteins' activities and, eventually, stopping the biological process that caused the illness (Galluzzi et al., 2024). Inhibition of the same target protein in the heart, though, can have a major adverse effect (Kitakata et al., 2022). Among these, the buildup of unfolded or misfolded proteins caused by drugs poses a major obstacle to comprehending and treating cardiac dysfunction.

Urea and guanidine chloride (GdmCl) have been widely utilized as chemical denaturants in both research and industrial facilities for an extended period of time (Paladino et al., 2023). A urea molecule forms five interactions with the protein moiety, while a Gdm + ion forms six interactions with the protein moiety and has a notable affinity for aromatic side chains (Paladino et al., 2022). Their denaturing impact is based on their heterophilicity. The primary factor that causes unfolding is the rise in entropy resulting from the generation of a vast number of conformational microstates due to the occupation and non-occupation of the binding site (Seelig and Seelig, 2023). Interestingly, urea-based anticancer agents have grown by leaps and bounds over the past few decades (Listro et al., 2022), especially a series of protein kinase inhibitors including sorafenib (Wilhelm et al., 2006) and regorafenib (Dhillon, 2018). The urea moiety is supposed to be introduced to serve as a favorable scaffold and to enhance affinity with the substrate (Ghosh and Brindisi, 2020), but such drugs also have the ability to act as in vivo denaturants, and therefore tight blood concentration monitoring is necessary.

In addition to the direct action of denaturants, there are multiple mechanisms that lead to disequilibrium in proteostasis, including ER stress, unfolded protein response (UPR), chaperone dysfunction, impairment of the proteasome system, and dysregulation of autophagy (Henning and Brundel, 2017; Ren et al., 2021). ER stress and UPR, initiated by the accumulation of misfolded proteins within the ER, trigger a cascade of signaling events aimed at restoring protein homeostasis but can ultimately induce apoptosis and lead to cardiac dysfunction if unresolved (Foufelle and Fromenty, 2016). Uncontrolled ER stress and UPR induce apoptosis by a variety of pathways, such as the promotion of C/EBP homologous protein (CHOP) expression by activating transcription factor 4 (ATF4), inhibiting the B cell lymphoma protein 2 (BCL-2) protein family to induce apoptosis (Ren et al., 2021). Inositol-requiring enzyme-1 (IRE-1) can also form a complex with tumor necrosis factor receptor-associated factor-2 (TRAF2) and apoptosis signal-regulating kinase-1 (ASK1), activating c-Jun N-terminal kinase (JNK) signaling as well as caspase-12 (Sano and Reed, 2013). The activating transcription factor 6α (ATF6α) can also induce apoptosis by regulating CHOP expression or downregulating myeloid cell leukemia-1 (MCL-1) (Adachi et al., 2008; Fernández et al., 2015). Chaperones play a crucial role in assisting protein folding and preventing aggregation. Meanwhile, Hsp 70, a major category of chaperons, can directly interact with apoptotic protease-activating factor 1 (Apaf-1) which is a key regulator of the caspase-dependent apoptotic pathway and prevents oligomerization of Apaf-1 with procaspase-9, thereby inhibiting apoptosome formation (Ravagnan et al., 2001). However, the expression and function of the chaperones can be compromised by drug-induced stressors, which can lead to uncontrolled ER stress and apoptosis (Mayer and Bukau, 2005; Fu et al., 2016). Additionally, impairment of the proteasome system, responsible for degrading misfolded proteins, further exacerbates protein accumulation within cardiac cells (Li et al., 2020; Zheng et al., 2023). These abnormal ubiquitinated protein aggregates can induce cellular injury and ultimately lead to caspase-mediated apoptosis and cell death (Hasinoff et al., 2017). Moreover, dysregulation of autophagy, a vital cellular process for degrading protein aggregates, can contribute to the accumulation of toxic protein species and subsequent cardiac dysfunction (Kim et al., 2011; Nazarko and Zhong, 2013; Li et al., 2020) (Figure 2).

Understanding the intricate interplay between these pathways is essential for developing targeted therapeutic strategies to alleviate drug-induced cardiac dysfunction (Table 2).

In comparison to other cell types, cardiomyocytes are of particular necessity to regulate proteostasis and are therefore susceptible to the side-effects of drugs due to their highly specific functions (electrical conduction, contraction), high metabolic demands, and terminally differentiated state (Henning and Brundel, 2017). At present, three categories of primary prevention measures for DICT are in use (Curigliano et al., 2016). The most common is dose restriction, but this is likely to affect antitumor therapy outcomes. Furthermore, there is no absolute safe dose for a particular drug due to individual differences. The second category comprises the use of alternative treatments or changes in the mode of administration. The third category is the combination of cardioprotective drugs, such as atorvastatin and sacubitril/valsartan (Li et al., 2022), but their actual utility has yet to be verified in further basic and clinical trials.

The only currently approved pharmaceutical agent for the prevention of anthracycline-induced cardiotoxicity is dexrazoxane (Langer, 2014). However, the scenarios for its use are limited, thereby necessitating the development of novel drug therapies. The mechanism by which the aforementioned pharmaceutical agents induce cardiotoxicity by disrupting proteostasis presents a promising avenue for mitigating DICT. Geranylgeranylacetone (GGA) is the most intensively studied of the compounds under consideration. It is not only a widely used antiulcer drug but also a non-toxic inducer of HSP expression. It has been demonstrated to rapidly increase the expression of HSPs in response to ischemia, hypoxia, oxidative stress, and toxicity, resulting in a significant protective effect. Its efficacy has been established in the prevention and treatment of myocardial ischemia-reperfusion injury and atrial fibrillation (Zeng et al., 2018). Notably, in doxorubicin-induced cardiotoxicity, GGA could inhibit oxidative stress and apoptosis in cardiomyocytes by upregulating HSP90 (Fujimura et al., 2012; Sysa-Shah et al., 2014). Similarly, Salubrinal upregulated the chaperone levels in cardiomyocytes via suppressing the dephosphorylation of eIF2A, thus protecting against ER-stress induced apoptosis (Li et al., 2015). Furthermore, a recently published study has confirmed a novel mechanism by which low-dose colchicine attenuates doxorubicin-induced cardiotoxicity by promoting autolysosome degradation through microtubule regulation (Peng et al., 2024). Given its low cost, long-term safety, and good tolerability, colchicine has the potential to be an optimally suited drug candidate. The subsequent phase of research should concentrate on the cardiac-specific regulation of proteostasis, to develop innovative drug therapies that can mitigate cardiotoxicity without affecting the efficacy of the chemotherapeutic agents themselves.

The prevailing methodologies for evaluating drug cardiotoxicity concentrate on cumulative dose exposure to drugs and the monitoring of cardiac function (ultrasound, electrocardiogram, etc.), yet they are deficient in terms of temporal precision and specificity. Technological advances and methodological breakthroughs in recent years have enhanced our comprehension of protein misfolding. The monitoring of protein homeostasis in clinical samples may prove to be a promising avenue for effectively assessing potential drug cardiotoxicity. Furthermore, advancements in preclinical trial modeling are occurring simultaneously with the monitoring of protein homeostasis (Lee et al., 2024). Among them, human induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) are a highly effective platform (Burnett et al., 2021). Adult CMs and iPSC-CMs exhibit a significant degree of similarity in terms of their electrophysiology, cellular signaling, contractile processes, and metabolism (Lindenstrauß et al., 2017). The intricate and coordinated functions have direct relevance to the assessment of toxicity.

As mentioned previously, several spectroscopic methods have been used to directly monitor the structural changes and protein conformation. Our current priority is to explore the practical application of these approaches in the in vivo setting (Doerr, 2014). LiP-MS, or Limited Protein Hydrolysis Coupled Mass Spectrometry, is a newly developed proteomics technique that can detect alterations in protein structure within intricate biological settings, covering the entire proteome (Cappelletti et al., 2021). The quantification of conformationally specific peptides produced by protease digestion allows for the study of the structure of disease-related proteins in clinical samples, enabling the selection of potential biomarkers. LiP-MS screening of proteomic changes in cerebrospinal fluid (CSF) identified structural changes in proteins in the CSF that are closely associated with neurodegenerative pathologies (Shuken et al., 2022). Furthermore, LiP-MS analysis of samples in the presence of drug candidates permits a direct assessment of the effect of drug candidates on aggregation processes (Holfeld et al., 2024). Further, Currie et al. developed an experimental strategy and computational analysis workflow to perform simultaneous proteome localization and turnover (SPLAT) measurements (Currie et al., 2024). The spatiotemporal proteomic data obtained from carfilzomib-intervened IPSC-CM revealed significant disturbances in proteasome remodeling, the induction of chaperone and ERAD proteins, and the mitochondrial protein quality control mechanisms. Additionally, a preferential decrease in sarcomere indicates that it may be the primary site of lesion for carfilzomib cardiotoxicity. Furthermore, differential protein abundance analysis revealed significant upregulation of MHC-β (MYH7) and desmoplakin (DSP), suggesting that similar protein deposition lesions may underlie the mechanism of cardiotoxicity in vivo.

In addition, thermal proteome profiling (TPP), which combines cellular thermal shift assay (CETSA) and quantitative mass spectrometry (MS), can depict the melting profile of thousands of proteins. With the addition of a specific drug, the melting curves are evaluated for differences, thus the target proteins were identified (Saei et al., 2020). Perrin et al. exploited tissue-TPP and blood-TPP techniques to reveal the proteome-wide heat stability profiles of the deacetylase inhibitor panobinostat, as well as the B-Raf inhibitor vemurafenib, and to derive their target profiles (Perrin et al., 2020). These approaches will help to elucidate the mechanism of drug action in vivo and make systematic assessments of the stability of target proteins.

Given that these proteostructural monitoring methods have only recently emerged, the related basic experiments and clinical studies are still in their infancy. Fortunately, preliminary basic studies have demonstrated the unique value of these methods in elucidating protein homeostasis, including target identification, spatial localization, and quantitative analysis. We anticipate that the relevant technologies will prove invaluable in the subsequent development of new drugs or monitoring of drug efficacy.

The research about the dynamics of proteostasis is going through a paradigm shift. Increasingly more people are emphasizing the need to establish an entire in vivo research system. Relevant research should concentrate on the dynamic monitoring of proteostasis and its control systems, particularly chaperones, proteasomes and others. It can be concluded that antineoplastic drugs have the potential to disrupt cardiac proteostasis through four mechanisms, including the induction of ER stress and UPR, the interference with the expression or function of chaperones, the alteration of proteasome function, and the influence of autophagy, which can result in cardiotoxicity such as cardiomyopathy, myocarditis, and arrhythmia. This represents a novel avenue for the prevention and treatment of DICT, whereby pharmacological agents targeting proteostasis, including GGA, can be employed to mitigate proteotoxicity. The advent of novel in vivo protein monitoring techniques, such as LiP-MS, has the potential to furnish more accurate insights into proteostasis. What’s more, they also offer a novel approach to future drug development, whereby in vivo model-based high-throughput screening techniques are employed to identify pharmacological targets, thus mitigating the risk of cardiotoxicity.