The Role of Nrf2 and Inflammation on the Dissimilar Cardiotoxicity of Doxorubicin in Two-Time Points: a Cardio-Oncology In Vivo Study Through Time

Doxorubicin (DOX) is a DNA topoisomerase II inhibitor used in several types of cancers [1‐3]. DOX is a powerful anticancer drug, but its use is associated with significant adverse side effects. Cardiotoxicity is one of the most critical adverse effect of DOX and the second cause of death among cancer survivors. Clinically, cardiotoxicity triggered by DOX is manifested by arrhythmias, systolic dysfunction, and heart failure (HF), among other effects [1, 2]. DOX can cause cardiotoxicity early in the treatment (up to 2 weeks after its end), being named acute/subacute cardiotoxicity. The clinical manifestations of this toxicity usually manifest themselves as chest pain, palpitation, dysplasia, and/or tachycardia arrhythmias, as well as a decline in the left ventricle ejection fraction (LVEF) [1, 2, 4]. When the effects manifest later (weeks to years after treatment has ceased), it is called chronic cardiotoxicity and is characterized by dilated cardiomyopathy, including dilation of ventricles, in some cases of atria, reduced LVEF and contractile function, diastolic dysfunction, and mural thrombi in some patients [1, 2, 4]. The incidence of late cardiotoxicity exponentially increases after the cumulative dose of 400–550 mg/m² [1, 2], and DOX-induced cardiotoxicity is considered multifactorial. Several mechanisms have been suggested over the years to explain the cardiotoxicity caused by DOX: reactive oxygen species (ROS) formation has been the most reported and considered the primary mechanism of cardiotoxicity, as the chemical reactivity of DOX involves a redox cycle that results in the generation of ROS [5, 6]. On the other hand, DOX may also directly bind to endothelial nitric oxide (^•NO) synthase and generate a DOX-semiquinone radical [7]. Mitochondrial DNA lesions caused by ROS or directly by DOX can lead to respiratory chain failure and ROS formation [8]. Nonetheless, antioxidants do not revert the inflicted damage when experimental and clinical trials address the long-term effects of the drug [9, 10]. Other contributors to DOX-induced cardiotoxicity include inhibition of nucleic acid and protein synthesis, the release of vasoactive substances, disturbed function of myocardial adrenergic receptors and adenylate cyclase activity, alterations in lysosomal morphology and enzyme activities, disruption of calcium transport in the cardiac sarcolemma, accumulation of iron in mitochondria, the formation of cardiotoxic metabolites and the activation of apoptosis [11, 12]. However, extensive research has revealed that DOX may unleash alternative cardiac damaging mechanisms, such as cardiac topoisomerase 2β inhibition, autophagy, pyroptosis, immunometabolism, and inflammation [13‐16].

Moreover, the classic vision has been that DOX-induced cardiac inflammation occurs due to the production of ROS and the subsequent apoptosis of cardiomyocytes [17], which provokes an increase in pro-inflammatory cytokines, inflammatory cell infiltration, and necrosis in cardiac tissues [18]. Accordingly, in our previous in vivo work, which used clinically relevant doses, we showed that adult mice seem to be more prone to DOX-induced cardiotoxicity, by mechanisms related to inflammation when compared with infant mice [19]. Several in vivo studies using short-term administrations have demonstrated that DOX leads to inflammation of the cardiac tissue [20‐22]. However, as far as we know, the role of inflammation in the long-term cardiotoxicity of DOX has yet to be determined.

Moreover, we demonstrated that DOX causes a cardiac inflammatory response and changes the hearts’ redox response [19]. Inflammation is closely linked to cardiac redox homeostasis, being that the transcription factor nuclear factor erythroid 2-related factor 2 (Nrf2) is a paramount link. Nrf2 is a regulator of multiple cytoprotective genes, which maintain redox homeostasis and exert anti-inflammatory functions; thus, it is a strong contributor to cardiovascular health [23]. Over the last few years, several studies have supported the regulation of DOX cardiotoxicity by Nrf2 [24‐26]. The activation of the Keap1/Nrf2 antioxidant response system has been identified as an important cellular defense mechanism against oxidative stress after acute exposure to DOX in vitro [27]. Among the antioxidant enzyme expression regulated by Nrf2 are catalase, superoxide dismutase, glutathione S-transferases, and glutathione peroxidases, among other redox regulators [23, 28]. In addition, the Nrf2 signaling pathway exerts a negative regulatory influence on various inflammatory mediators such as cytokines, chemokine-releasing factors, matrix metalloproteinases, cyclooxygenase-2 (COX-2), and inducible nitric oxide synthase (iNOS). These mediators, in turn, directly or indirectly impact inflammation-controlling networks, including nuclear factor κB (NF-κB) and mitogen-activated protein kinase (MAPK) pathways, among others [28, 29]. Therefore, and considering the existing knowledge, this study aimed to evaluate the role of inflammation in DOX-induced cardiotoxicity and its possible link to its underlying late cardiotoxicity by using a clinically relevant dose in adult mice. Hence, Nrf2, NF-κB, and other redox and inflammation responders were assessed herein to gain a deeper disclose of the pathways involved in DOX-inflicted cardiotoxicity, mainly in the long-term. As far as we know, no other pre-clinical study evaluated the impact of DOX 5 months after the last administration.

Doxorubicin hydrochloride (≥ 98% purity, DOX), Ponceau S, direct red 80, 5,5-dithiobis(2-nitrobenzoic acid), adenosine triphosphate (ATP), reduced glutathione (GSH), glutathione reductase, oxidized glutathione (GSSG) disodium salt, bovine serum albumin, and the all other chemicals used were purchased from Sigma-Aldrich (St. Louis, MO, USA). Phosphate-buffered saline solution was purchased from Biochrom (Berlin, Germany), and sodium chloride (NaCl) was acquired from VWR (Leuven, Belgium). Isoflurane (Isoflo^®) was obtained from Abbott Animal Health (North Chicago, IL, USA). Harris haematoxylin was purchased from Harris Surgipath (Richmond, IL, USA), and 1% aqueous eosin from Australian Biostain (Traralgon, Australia). The Bio-Rad DC protein assay kit was obtained from Bio-Rad Laboratories (Hercules, CA, USA). Primary antibodies were acquired from different sources which are indicated in Supplementary Table S1. Goat anti-rabbit IgG-horseradish peroxidase (ab97051) and goat anti-mouse IgG-horseradish peroxidase (ab6728) were provided by Abcam (Cambridge, UK), while enhanced chemiluminescence (Clarity Western ECL, 1.705.060) reagents and the Bio-Rad DC protein assay kit were purchased from Bio-Rad Laboratories (Hercules, CA, USA). Amersham Protran nitrocellulose blotting membranes (0.45 µm) were supplied by Cytiva (Buckinghamshire, UK).

This work is the first to make a study of the impact of DOX in CD-1 male mice 5 months after the last administration. The major findings of this work were (1) even 5 months after DOX treatment, the body weight was still affected in mice; (2) histopathological examination allowed to see cardiac injury at 1W after DOX; (3) both DOX-treated groups had a significant increase in fibrotic tissue compared to control; (4) glutathione peroxidase, catalase, and iNOS expression increased in the 1W-DOX group, while in the 5M-DOX group, an increase in SOD2 expression and a tendency for the increase of catalase expression were seen; (5) the 1W-DOX group, a significant decrease in Nrf2 expression and a significant increase in p62 expression were observed, while in the 5M-DOX group, a significant increase in Nrf2 expression was seen; (6) in the 1W-DOX group, a tendency for an increase in Bax expression was seen, while in the 5M-DOX group, a significant decrease in Bcl-2 expression and an increase in the ratio of Bax/Bcl-2 expression were found. Regarding markers of inflammation: (1) in both DOX-treated groups, a higher density of infiltrating M1 macrophages was seen, although only the 1W-DOX group had a higher number of NF-κB p65 immunopositive cells in the cardiac tissue; (2) in the 1W-DOX group, a significant increase in TNFR2 expression and a tendency for increased TNFR1 and IL-6 expression were observed; (3) in the 5M-DOX group, a significant decrease of heart COX-2 expression, a tendency for decreased p38 MAPK expression, and a significant increase of myeloperoxidase, IL-33, and TNF-α expression were detected.

While in the 1W-DOX group, the body weight was constant throughout the experiment after 9.0 mg/kg of DOX, in the animals 5M-DOX, there was a significant body weight decrease in the last days of the protocol. To the best of our knowledge, no long-term study like this was previously done and we demonstrated that although a low cumulative dose (corresponding to 54.45 mg/m² in humans) has been used, DOX has a strong impact months after administration.

In the present work, we observed that even at a 9 mg/kg cumulative dose of DOX, cardiac tissue lesions were seen. Similar results were reported in our previous work [19] and by other authors [37‐45], when higher cumulative doses were used, revealing cellular damage, mostly swollen and vacuolated cardiomyocytes, capillary congestion, interstitial oedema, disorganization of myofibrils, and interstitial inflammatory cell infiltration, in different degrees. On the other hand, a semi-quantitative histopathologic analysis of the heart showed that it seems the heart had partially recovered, at least when assessing the numerical values of the semi-quantitative analysis from DOX treatment at 5M. Nevertheless, we observed that DOX treatment induced a significant increase in collagen deposition in the cardiac tissue of mice sacrificed 1W and 5M after the last administration. This result is according to what has been previously published in animal models [19, 46, 47].

Oxidative stress is the most frequently proposed mechanism to explain the complex pathophysiology of DOX-induced cardiotoxicity [2]. The heart is very prone to oxidative damage due to low levels of antioxidant enzymes. Although some studies have shown that the treatment of animals with antioxidants protects the heart against the toxicity of DOX [48, 49], no antioxidants have been proven to have clinical efficacy [50]. Conversely, the literature has descriptions that overexpression of antioxidant enzymes such as SOD2, catalase, or glutathione peroxidase in cardiomyocytes of transgenic mice attenuates DOX-induced cardiac damage [51‐53]. In the present work, DOX did not alter SOD2 expression but significantly increased catalase and glutathione peroxidase expression in the heart of the 1W-DOX group. Interestingly, DOX-induced contractile and mitochondrial dysfunction in the heart of mice were prevented by the overexpression of glutathione peroxidase [53] and what we see in our present work may be an adaptative response to cope with DOX-inflicted cardiotoxicity. In the 5M-DOX group, only an increase in SOD2 expression and a tendency to increase catalase expression were observed. Increased SOD2 expression has been shown to protect mitochondria from oxidative damage, decrease apoptosis, and preserve left ventricular function [54, 55]. Overall, our results suggest that DOX induces different responses regarding redox homeostasis. Nonetheless, we observed that protein oxidation by carbonylation was not seen in the 1W-DOX group, possibly because of the lower cumulative dose, when compared to previous studies [19, 37, 38, 56, 57], in which increased cardiac protein carbonylation was seen. Conversely, the 5M-DOX group had decreased carbonylated protein expression, showing that repair responses were triggered as seen in our previous study [19]. Possible, activation of the proteasome system acting on oxidatively modified proteins or other mechanisms may be involved that still need to be clarified.

iNOS is expressed in the heart upon inflammatory stimuli to engage in the production of excessive amounts of ^•NO [58] and it is considered to be a marker for oxidative/nitrosative stress [59]. In studies with mice that received DOX (20 mg/kg ip), a marked reduction in cardiac contractility was observed 5 days after administration, associated with a significant increase in myocardial iNOS immunopositivity and 3-nitrotyrosine formation [60, 61]. In addition, Mukhopadhyay and co-workers observed an increase in myocardial iNOS expression after DOX (20 mg/kg body weight ip) administration that peaked around day 5 when myocardial dysfunction was evident [62]. Moreover, the literature shows that iNOS induction plays a pathogenetic role in the development of chronic DOX-induced HF [63]. However, other studies have shown that iNOS-derived ^•NO may have protective effects on the heart, such as reducing myocardial damage after ischemia–reperfusion injury [64]. Although contradictory data may arise from other works, herein iNOS expression increased in the mice heart after DOX (1W-DOX group), and after the initial increase of iNOS expression, a normalization of response was seen, and no meaningful changes were observed in iNOS expression in the 5M-DOX group as compared with controls. Overall, the role of iNOS in the heart is complex and context-dependent, and we cannot take robust conclusions on its role on the cardiotoxicity seen here.

While oxidative stress has been linked to DOX cardiotoxicity, one of the key players of redox homeostasis is Nrf2 [63]. ROS generation is counteracted by cytoprotective mechanisms via the regulation of KEAP1/Nrf2 signaling [65]. Under physiological conditions, and through the canonical KEAP1/Nrf2 pathway, Nrf2 is suppressed by the negative regulator KEAP1, which leads to its ubiquitylation and proteasomal degradation. On the other hand, when Nrf2 escapes from the KEAP1 interaction, it translocates into the nucleus, where it regulates the expression of antioxidant and anti-inflammatory genes [66]. Nonetheless, several non-canonical pathways for Nrf2 activation engage competitive inhibition of the KEAP1/Nrf2 interaction by intracellular proteins such as p62 [67]. p62 is involved in various cellular processes, including autophagy, and signaling pathways related to inflammation and oxidative stress [68]. p62 sequesters KEAP1 within autophagosomes, preventing ubiquitylation of Nrf2 and leading to the release of Nrf2 into the nucleus. Therefore, the p62-KEAP1-Nrf2 signaling pathway plays an important role in the cellular response to oxidative stress [69, 70]. In this work, in the 1W-DOX group, a significant increase in p62 expression was observed; however, Nrf2 is contra-intuitively decreased. Decreased Nrf2 expression could potentially limit its ability to counteract oxidative stress and inflammation [71]. Herein, p62 does not seem to directly affect Nrf2, being possibly involved in other pathways e.g. autophagy. A study by Li and collaborators showed that Nrf2 deficiency exacerbates DOX-induced cardiotoxicity and cardiac dysfunction, suggesting that Nrf2 could be an endogenous suppressor of DOX-induced cardiotoxicity by controlling both oxidative stress and autophagy in the heart [24]. On the other hand, in the present work in the 5M-DOX group, a significant increase in Nrf2 expression was seen, corroborating a fully active proteasome in this time course. Moreover, while Bax was increased in the short term, its levels have normalized in the long term, indicating a potential shift in the regulation of cell survival and apoptosis. Although other mechanisms can be involved, in the 5M-DOX group (long-term), the activation of Nrf2 seems to be cardioprotective and interacts with apoptotic genes and possibly contributing to the damage recovery.

Nrf2 and NF-κB are key pathways regulating the balance of cellular redox status and responses to stress and inflammation [72]. NF-κB activity is limited when Nrf2 activation occurs [28]. In contrast, the inactivation of Nrf2 leads to the loss of NF-κB suppression and therefore upregulation of inflammatory responses [28]. In the 1W-DOX group, a higher density of infiltrating M1 macrophages and a higher number of NF-κB p65 immunopositive cells were seen with lower Nrf2 expression. M1 macrophages are required for the induction of a large number of inflammatory genes, including TNF-α, IL-1β, IL-6, IL-12p40, and COX-2, involved in various inflammatory processes [73], which may result in cardiac damage. The literature shows that the levels of pro-inflammatory cytokines, IL-1β [20, 41, 74], IL-6 [74], and TNF-α [41, 74] are increased after DOX administration (mostly after higher cumulative doses than ours study), and those are implicated in cardiac pathogenesis and apoptosis. The 1W-DOX mice showed a significant increase in TNFR2 expression and a tendency for an increase in TNFR1 and IL-6 expression. TNFR1 promotes biological responses ranging from NF-κB activation to cell death [75, 76], corroborating our data of NF-κB and Bax described earlier. In our previous work, DOX-treated adult mice (18.0 mg/kg cumulative dose) showed a tendency towards increased TNFR2, perhaps because this receptor activates NF-κB [19]. On the other hand, the heart was able to cope with the initial inflammatory stimuli caused by DOX, seen by increased expression of TNFR2, and a tendency for increased TNFR1 and IL-6 expression. However, in the 5M-DOX group, no meaningful differences in NF-κB p65, IL-1β, IL-6, TNFR1, and TNFR2 expression were seen. In fact, although apparently contra-intuitive in the 5M-DOX group, an increase in M1 macrophages was seen, while no changes in NF-κB p65 immunopositive were detected, and simultaneously Nrf2 increased expression. Thus, we hypothesize that DOX may have triggered, even at this low cumulative dose, inflammatory mediators, and some persisted throughout 5 months but were mainly abrogated (with the exception of M1 cells and related myeloperoxidase, TNF-α and IL-33). IL-33 is a cytokine produced by cardiac fibroblasts that can cause cardioprotective effects against hypertrophic remodeling and myocardial fibrosis by antagonizing angiotensin-II signaling and promoting anti-apoptotic factors, respectively [77, 78]. Moreover, in a model of recurrent neonatal seizure, IL-33 provided protection by suppressing apoptosis, and NF-κB-mediated inflammatory pathways, while maintaining p62 in normal levels [79].

Recent studies have shown that COX-2 and Nrf2 may interact to regulate inflammation and oxidative stress in the heart. COX-2 is an enzyme that is released at the site of tissue injury to produce prostaglandin E2 that stimulates inflammation. COX-2-derived prostaglandins have been shown to inhibit Nrf2 activity, leading to increased oxidative stress and inflammation in the heart [80, 81]. In contrast, Nrf2 activation in more severe and prolonged states of oxidative stress has been shown to inhibit COX-2 expression and inflammatory mediators [82], leading to reduced inflammation and oxidative stress. In fact, at 5M we see that an increase in Nrf2 expression is accompanied by COX-2 levels decrease. In the present work, no changes were seen in COX-2 expression in the heart at the earlier time point, but it decreased in DOX animals after 5M. Saito et al. have suggested that the inhibition of COX-2 with a selective COX-2 inhibitor improves cardiac function when given after the period of acute infarction in some experimental rodent models [83]. Another study showed that a selective COX-2 inhibitor given before and during the period of acute infarction reduced macrophage infiltration and fibroblast proliferation in rats’ hearts [84]. Scheuren et al. [84] suggested that angiotensin-II regulates COX-2 expression through the p38 MAPK pathway in isolated cardiac fibroblasts. We observed that DOX-treated mice (5M-DOX) had a tendency towards decreased p38 MAPK expression, which may be correlated with the decreased expression of COX-2. Overall, the results here presented suggest that the decrease of COX-2 at DOX-5M could help mitigate injury and inflammatory responses, and it is suggestive of adaptation of the heart to DOX-induced cardiotoxicity over time. The results of this study represent new insights into the effects of DOX-induced cardiotoxicity at different time points that could help healthcare professionals address patient care. It also is important for healthcare providers to consider several factors, such as hypertension, diabetes, coronary artery disease, HF, and smoking, as risk factors for DOX-induced cardiotoxicity as well as concomitant pharmacotherapy [1] when treating patients with DOX. Only the full knowledge of the patients will allow them to take the right steps to minimize the risk of cardiotoxicity. Several modulators of the inflammatory response such as non-steroidal anti-inflammatory drugs, glucocorticoids, natural products [85], omega-3 fatty acids, and probiotics [86] can modify the effects of DOX treatment and affect the prognosis.

Our study demonstrated that, in the short-term evaluation, DOX induces pathways that generate adverse outcomes related to inflammatory processes and response to oxidative stress. However, months after the last administration, the heart activates other response mechanisms. Nonetheless, some inflammatory mediators continued over time, and we cannot ignore possible chronic inflammation. In the heart of the 1W-DOX group, a significant decrease in Nrf2 expression was observed, a significant increase in p62 expression, and a tendency for an increase of Bax, suggesting that the apoptotic route is initiated. A higher number of NF-κB p65 immunopositive cells may suggest an interaction between these factors in the cellular response to oxidative stress and inflammation, with Nrf2 blockage and activation of the apoptosis pathway. On the other hand, in the 5M-DOX group, a significant increase in Nrf2 expression was found, and normalization of Bax expression, although M1 cells are still present and possibly are still removing debris of apoptotic cells of the earlier damage. These results suggest that in the 5M-DOX group, the activation of Nrf2 seems to be cardioprotective. Moreover, in the 5M-DOX group, DOX led to a reduction in COX-2 expression and an increase in IL-33 expression, which may contribute to mitigating fibrosis and inflammatory responses. This DOX cumulative dose may have led to an imbalance on redox defenses earlier on, possibly related to inflammation, which activates Nrf2. However, it is important to note that changes in Nrf2 expression, p62 expression, and related factors are likely reactive responses of the organism to the initial damage induced by DOX. Furthermore, we should consider the possibility that inflammation may result from the initial damage caused by cardiotoxicity, rather than being the primary instigator of cardiotoxicity itself. We believe that the present study provides valuable information on the potential mechanisms underlying DOX-induced cardiotoxicity; however, we recognize several limitations to the work that deserve consideration, namely the assessment of cardiotoxicity in this study is mainly based on semi-quantitative histopathology and no functional assessment of heart function was done. Moreover, although we hypothesize that inflammation and the Nrf2 modulation can be key to the cardiotoxicity and pathways changes observed, one cannot overrule other molecular pathways not evaluated here as important also for the crosstalk that leads to DOX-induced cardiotoxicity.