Doxorubicin triggers splenic contraction and irreversible dysregulation of COX and LOX that alters the inflammation-resolution program in the myocardium

Abstract

Doxorubicin (DOX) is a widely used drug for cancer treatment as a chemotherapeutic agent. However, the cellular and integrative mechanism of DOX-induced immunometabolism is unclear. Two-month-old male C57BL/6J mice were divided into high- and low-dose DOX-treated groups with a maintained saline control group. The first group was injected with a high dose of DOX (H-DOX; 15 mg·kg⁻¹·wk⁻¹), and the second group was injected with 7.5 mg·kg⁻¹·wk⁻¹ as a latent low dose of DOX (LL-DOX). H-DOX treatment led to complete mortality in 2 wk and 70% survival in the LL-DOX group compared with the saline control group. Therefore, an additional group of mice was injected with an acute high dose of DOX (AH-DOX) and euthanized at 24 h to compare with LL-DOX and saline control groups. The LL-DOX and AH-DOX groups showed obvious apoptosis and dysfunctional and structural changes in cardiac tissue. Splenic contraction was evident in AH-DOX- and LL-DOX-treated mice, indicating the systems-wide impact of DOX on integrative organs of the spleen, which is essential for cardiac homeostasis and repair. DOX dysregulated splenic-enriched immune-sensitive lipoxygenase and cyclooxygenase in the spleen and left ventricle compared with the saline control group. As a result, lipoxygenase-dependent D- and E-series resolvin precursors, such as 16HDoHE, 4HDoHE, and 12-HEPE, as well as cyclooxygenase-mediated PG species (PGD₂, PGE_2, and 6-keto-PG_2α) were decreased in the left ventricle, suggestive of defective immunometabolism. Both AH-DOX and LL-DOX induced splenic contraction and expansion of red pulp with decreased CD169⁺ metallophilic macrophages. AH-DOX intoxicated macrophages in the spleen by depleting CD169⁺ cells in the acute setting and sustained the splenic macrophage loss in the chronic phase in the LL-DOX group. Thus, DOX triggers a vicious cycle of splenocardiac cachexia to facilitate defective immunometabolism and irreversible macrophage toxicity and thereby impaired the inflammation-resolution program.

NEW & NOTEWORTHY Doxorubicin (DOX) triggered splenic mass loss and decreased CD169 with germinal center contraction in acute and chronic exposure. Cardiac toxicity of DOX is marked with dysregulation of immunometabolism and thereby impaired resolution of inflammation. DOX suppressed physiological levels of cytokines and chemokines with signs of splenocardiac cachexia.

INTRODUCTION

The anthracycline-based cancer drug doxorubicin (DOX; also known as adriamycin) is a chemotherapeutic agent that has been routinely used in wide variety of cancers, such breast, ovarian, bladder, lung, thyroid, and stomach cancers (4, 23, 29). Previous studies have suggested that DOX is used in 50%–60% of breast cancer and 70% of childhood cancer treatment protocols (19). Despite these records, treatment with DOX has been reported to cause dose-dependent cardiac toxicity and heart failure (18, 39). Use of DOX results in progressive cardiotoxicity and can lead to severe myocardial damage (25, 29). However, the cellular and integrative mechanism of the underlying acute and long-term DOX-induced myocardium pathology is unclear. Many studies have indicated that generalized oxidative stress and myocardium apoptosis play an important role in determining the fate of cardiomyocytes (15, 16, 36, 42). Thus, acute DOX exposure or long-term risk after initial drug exposure is of great importance to cancer survivors and the understanding of the cellular and molecular mechanism in DOX-induced cardiovascular pathobiology (28).

DOX-mediated cardiac toxicity begins within hours or weeks and is progressive after years of treatment. The risk of DOX-induced heart failure almost doubles with advanced aging. Retrospective trials have shown that the DOX effect is dose and schedule dependent; almost 18% of patients show signs of cardiac toxicity with a high dose (34, 37). The mechanisms for acute DOX-induced adverse changes in cardiac pathology are unclear. DOX-induced long-term effects due to cumulative drug exposure are of much concern, such as acute dose incidents (31). DOX-induced cardiac pathology appears within 30 days of administration, but it may also appear 6–10 yr after DOX exposure in cancer survivors (18, 30, 39). Therefore, we precisely compared acute DOX exposure and compared it with long-term effects on the heart and spleen with an integrative impact on immunometabolism.

Immunometabolism is primarily driven by immune-sensitive enzymes, such as cyclooxygenases (COXs) and lipoxygenases (LOXs). COX-1 and COX-2 are isoforms of the COX family that use arachidonic acid as a substrate to form differential eicosanoids (27). Inhibition of COX-2 causes major adverse effects with an incidence of myocardial infarction and readmission because of heart failure (11). Thus, the homeostatic and cardioprotective role of COX-2 has been well described in a DOX-induced rat model (1). COX-2 is believed to be cardioprotective through the induction of molecules, such as prostacyclins (32). The administration of a COX-2 inhibitor prevents the enzyme from generating prostacyclins, eliminating the cardioprotective mechanism (24). LOXs are immune-responsive fatty acid metabolic enzymes that regulate cardiac and immune metabolism and biosynthesize D- and E-series resolvin species in coordination with the spleen (24). Here, we defined the cellular and molecular events of defective immunometabolism after acute and chronic DOX exposure using physiological and toxic doses. In this study, we established that DOX activated splenocardiac dysregulation with marked incoordination of immune-responsive COX and LOX metabolic enzymes and thereby impaired the inflammation-resolution program.

MATERIALS AND METHODS

Animal compliance.

All animal procedures were conducted according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committees at The University of Alabama at Birmingham.

DOX protocol and study design.

Two-month-old male C57BL/6J mice were used for the experiments. C57BL/6J mice were randomized into four subgroups (10 mice/group except for the acute-dose DOX group) with saline as a vehicle (0.9% NaCl in water) for the control group. DOX at a high dose (H-DOX group; 15 mg·kg⁻¹·wk⁻¹) was used for survival analysis, 7.5 mg·kg⁻¹·wk⁻¹ DOX was used as a latent low dose (LL-DOX group), and 15 mg·kg⁻¹·24 h⁻¹ DOX was used as an acute high dose (AH-DOX group). All doses of DOX were given to mice via intraperitoneal injection. Echocardiography for strain analysis was performed at two different time points of the study: day 0 and day 14, respectively (Fig. 1A). The second portion of the animal study was performed with acute-dose DOX (15 mg/kg) for 24 h in five experimental mice. For a translational perspective, 5 mg/kg DOX was used as a physiological dose for flow cytometry-based leukocyte quantitation of splenic mononuclear cells and compared with 7.5 mg/kg DOX in an acute setting. The mortality rate in each group was checked every day.

Fig. 1.

High-dose doxorubicin (H-DOX; 15 mg/kg) and latent low-dose (7.5 mg/kg) doxorubicin (LL-DOX) defined nonsurvival and survival doses in cardiac toxicity. A: study design illustrating the long-term study of DOX using male C57BL/6J mice to define the integrative mechanism of DOX on the spleen and heart. Cardiac function was measured on day 0 (baseline) and day 14 after treatment. B: scheme illustrating the acute study (24 h) of DOX treatment in C57BL/6J mice designed to compare with the low-dose chronic effect. C: survival rates in LL-DOX- and H-DOX-injected mice compared with saline control-injected mice and analyzed by a log-rank test. Survival was 10 of 10 (male) mice in the saline control group,  0 of 10 (male) mice in the H-DOX group, and 3 of 10 (male) mice in the LL-DOX group. *P < 0.01 vs. the saline control group. The pie chart represents survival and mortality percentages in the LL-DOX group on day 14. D: representative hematoxylin and eosin (H&E)-stained (left), picrosirius red (PSR)-stained (middle), and TUNEL-stained (right) left ventricular images at ×40 magnification accompanied with images at ×1.25 magnification from the saline control, acute high-dose DOX (AH-DOX), and LL-DOX groups. Scale bar = 50 μm. n = 6 mice/group. E: representative bar graph of the percentage of fibrosis from PSR-stained AH-DOX and LL-DOX groups compared with the saline control group. F: representative bar graph of the number of apoptotic cells from TUNEL-stained AH-DOX and LL-DOX groups compared with the saline control group. n = 2–3/group.

Transthoracic echocardiography.

The Vevo 3100 (VisualSonics) in vivo imaging system was used to perform echocardiography equipped with a probe up to 40 MHz with a resolution of 30 μm. Mice were anesthetized with 1.5–2% isoflurane in an oxygen mix. Heart rate (>400 beats/min), respiratory rate, and body temperature (35–37°C) were continuously monitored throughout the procedure to ensure an adequate depth of anesthesia. Echocardiographic speckle tracking-based strain measures of myocardial deformation were obtained from two-dimensional grayscale echocardiography images acquired from long- and short-axis views. Longitudinal strain was measured and reported for long-axis parameters (9).

Necropsy.

Mice were anesthetized with 2% isoflurane in an oxygen mix and injected with 4 IU/g heparin as previously reported. The lungs, left ventricle (LV), right ventricle, and spleen were separated and weighed individually. Half of the spleen and a middle section of the LV were either fixed in 10% zinc formalin for immunohistochemistry or kept in OCT for cryosectioning. The rest of the spleen and LV were snap frozen for molecular analysis as previously described (9).

LV and spleen hematoxylin and eosin staining.

For histological measurements, LV midcavity and spleen longitudinal sections were embedded in paraffin and sectioned. LV and spleen sections were stained with hematoxylin and eosin. LV and spleen images were acquired for each mouse using a microscope (BX43) with an attached camera (Olympus DP73). A total of five to six images were analyzed per mouse per group (9).

Collagen staining using picrosirius red.

Picrosirius red (PSR) staining was used to measure collagen density in paraffin-embedded LV midcavity sections of all experimental mice. Collagen density data were analyzed by software as previously described (9).

TUNEL assay.

Apoptotic cell death in cardiomyocytes in the midsection of the LV was detected by in situ TUNEL staining. TUNEL staining was performed on LV sections using the fluorescent in situ Cell Death Detection Kit (Promega G3250) according to the manufacturer's instructions. TUNEL-positive nuclei were counted per group by fluorescence microscopy, and an apoptotic index was determined as the percentage of TUNEL-positive nuclei, which was scored blindly by two evaluators.

Quantitative real time-PCR.

LV and spleen RNA was isolated from saline control, AH-DOX, and LL-DOX groups using the TRIzol method. For quantitative PCR, reverse transcription was performed with 2.5 μg total RNA using the SuperScript VILO cDNA Synthesis Kit (Invitrogen, Carlsbad, CA). Quantitative PCR for arachidonate 15-lipoxygenase (ALOX15; Mm00507789_m1), arachidonate 12-lipoxygenase (ALOX12; Mm00545833_m1), arachidonate 5-lipoxygenase (ALOX5; Mm01182747_m1), COX-1 (Mm00477214_m1), COX-2 (Mm00478374_m1), Bax (Mm00432051_m1), and Bcl2 (Mm00477631_m1) genes was performed using TaqMan probes (Applied Biosystems) on a master cycler ABI (7900HT). Gene expression was normalized to the housekeeping control gene hypoxanthine phosphoribosyltransferase-1 (HPRT-1; Mm01545399_m1). Results are reported as fold change values. All experiments were performed in duplicate with n = 4–6 mice/group.

LV protein extraction and immunoblot analysis.

LV tissues were processed for protein extraction as previously described (8). Electrophoresis of 10 µg LV protein was performed as previously described and probed with primary antibody (COX-1; 1:1,000) overnight at 4 °C followed by secondary antibody (Bio-Rad, Hercules, CA). Proteins were detected using the Femto chemiluminescence detection system (Pierce Chemical, Rockford, IL). Densitometry was performed using ImageJ software (National Institutes of Health) (20).

LV wheat germ agglutinin staining.

To analyze myocyte area, wheat germ agglutinin (WGA) staining was performed. Cryosections of the LV midcavity were fixed with 4% paraformaldehyde for 15 min and subsequently given three washes with PBS (5 min per wash). Sections were blocked with 10% goat serum (Vector-S-1000), and Alexa fluor 488-conjugated WGA (Invitrogen) solution was added to the tissue in 1:1,000 dilution for 1 h. Samples were then washed with PBS and mounted with prolonged gold antifade reagent (Invitrogen). Images were acquired using a Nikon A1R confocal microscope using a ×40 objective. Myocyte area was quantified from 5−6 high-power fields/section using ImageJ software (National Institutes of Health). Data for each group were calculated from 30 cardiomyocyte sections and 4–5 mice/group (12).

Confocal microscopy.

Immunofluorescence was performed as previously described (13). Briefly, spleens were fixed with OCT (Tissue TeK). For staining, LV midcavity and spleen tissues were fixed in 4% paraformaldehyde for 15 min and permeabilized in 0.1% Triton X-100 in PBS for 10 min. Tissues were blocked with 10% goat serum (Vector) and 1% BSA (ACROS) for 1 h. Samples were stained with F4/80 (1:100, Abcam) and CD169 (1:100, Novus) or α-smooth muscle actin (α-SMA, 1:1,000) antibodies overnight. Tissues were washed three times with PBS (10 min per wash) and probed with secondary anti-mouse Alexa 488/555 antibody (ThermoFisher Scientific, Waltham, MA). Nuclei were stained using Hoechst (1:2,000 in PBS) for 5 min. Tissues were washed three times with PBS as described above. Coverslips were mounted in prolong gold antifade reagent (ThermoFisher Scientific) and analyzed using a Nikon A1R confocal microscope. Images were acquired with a ×10 objective for the spleen and processed using a Nikon A1R confocal microscope.

Analysis of lipid mediators by mass spectrometry.

After necropsy, 15–20 mg of LV tissue from all three groups (saline, AH-DOX, and LL-DOX) were homogenized at a 1:9 ratio with ×1 PBS (pH 7.4) and centrifuged at 10,000 revolutions/min for 5 min at 4°C. The supernatant was collected, and protein was measured using a Bradford Kit (Bio-Rad). Lipid mediator (LM) was measured using liquid chromatography-tandem mass spectrometry (LC-MS/MS) as previously described (8, 20). For pie chart analysis, all LOX-mediated LMs were taken together, and the relative contribution of each LM was contributed to an overall total of LOX-mediated LMs.

Flow cytometry.

Single mononuclear cells were isolated from the spleen of DOX (5 and 7.5 mg/kg)-treated groups and compared with the saline-treated group. We used CD45-phycoerythrin-CY7 (BD Biosciences, San Jose, CA), CD11b-allophycocyanin, F4/80-PerCP, and CD169-Alexa fluor 488 (ThermoFisher Scientific) in a cocktail. The LIVE/DEAD Fixable Blue Dead Cell Stain Kit (ThermoFisher Scientific) was used to determine the viability of cells before fixation. The live cell population was primarily gated with CD45⁺ markers for hematopoietic cells. CD169⁺ macrophages were gated on CD11b (Mac-1) and F4/80⁺ cell surface markers. Cell count was measured on the LSRII Flow cytometer (BD Biosciences), and data were analyzed with FlowJo software (v.10.4.2, Ashland, OR).

Statistical analyses.

All data are expressed as means ± SE. Statistical analyses (P < 0.05) were performed using Graph Pad Prism 7. One-way ANOVA was used. The Kaplan-Meier test and log-rank test were followed for survival analysis. P < 0.05 was considered as statistically significant.

RESULTS

H-DOX and LL-DOX demarcate nonsurvival versus survival and cardiac toxicity.

DOX-induced heart failure has been extensively reported in humans (22, 38). To determine the cellular and integrative mechanism, nonsurvival H-DOX (15 mg·kg⁻¹·wk⁻¹) was used in the acute study (24 h) and LL-DOX was used in the chronic study (Fig. 1, A and B). AH-DOX was toxic, and all mice died within 2 wk, whereas LL-DOX showed 70% survival. The survival curve (Fig. 1C) of H-DOX (15 mg·kg⁻¹·wk⁻¹) showed 100% mortality within 13 days, and none of the mice survived at the end point of the study, indicating that H-DOX was a nonsurvival dose in C57BL/6J mice. Suboptimal LL-DOX (7.5 mg·kg⁻¹·wk⁻¹) led to 30% mortality (3 of 10 mice) in the first week of the treatment. The remaining 70% of survivors showed obvious cardiac toxicity signs with marked low body weight. These results demonstrate that LL-DOX was a survival dose compared with H-DOX-treated mice. DOX-induced mortality was accompanied by cachexic histological alterations in the myocardium, leading to cardiotoxicity. Furthermore, we examined the effect of AH-DOX and LL-DOX on the progression of cardiotoxicity in cardiac and spleen tissue sections. The disorganized myocardium was more apparent in LL-DOX-treated mice compared with saline control-treated mice. AH-DOX caused initiation of the perivascular cytoplasmic disorganization, indicating that cardiac pathology was the sign of structural tissue damage. Next, we evaluated cardiac fibrosis in all experimental animal groups by PSR staining and apoptosis with TUNEL assay in LV sections. As shown in Fig. 1D, LL-DOX treatment caused an increase in cardiac fibrosis. The significant increase in the percentage of cardiac fibrosis with an increase in an extracellular collagen matrix, which appeared red after PSR staining, was more predominant in LL-DOX mice compared with control mice (Fig. 1E). The degree of apoptosis was measured by the number of TUNEL-positive nuclei in cardiac tissue. AH-DOX- and LL-DOX-treated mice displayed an increase in apoptosis compared with saline control-treated mice (Fig. 1F). These results suggest that DOX-induced cardiac toxicity, as well as mortality, can be attributed to histological alterations in cardiac tissue, namely, an increase in fibrosis and apoptosis.

AH-DOX and LL-DOX induced cardiac dysfunction and lowered myocardium strain.

DOX-induced apoptosis and an increase in extracellular collagen formation were observed by histology. We further confirmed the fibrotic remodeling by α-SMA expression. Both AH-DOX- and LL-DOX-treated mice displayed higher α-SMA expression in the interstitial region compared with saline-injected control mice (Fig. 2A). The former results led to further investigation of cardiac function in AH-DOX and LL-DOX using echocardiography. Strain and synchronicity measurements allowed the precise measurement of ventricular function after acute and chronic exposure of AH-DOX and LL-DOX, respectively. Before DOX treatment, all three groups of mice showed typical fractional shortening (33–35%) and longitudinal strain of about −17 to −20, and all six regions were synchronized to maintain cardiac homeostasis. Both AH-DOX and LL-DOX groups showed decreased myocardium strain presented in midsystole with a marked decrease in green vector length, indicative of LV dysfunction (Fig. 2B). A globally reduced strain could be inferred through the decrease in strain values of systolic (red) and diastolic (blue) function in the three-dimensional strain as well as LV region dysfunction visualized through the decrease in synchronicity compared with the saline control group (Fig. 2C and Table 1). AH-DOX treatment facilitated cardiac dysfunction with a marked increase of end-diastolic dimension and end-systolic dimension compared with saline, as shown in Table 1. Function in the LL-DOX group recovered to some extent, but the dysfunction was sustained with some recovery compared with the AH-DOX group. Likewise, longitudinal strain and segmental synchronicity were both altered in LL-DOX. However, the dysfunction was permanent in the AH-DOX group (Fig. 2, C and D). Thus, the presented results show that AH-DOX and LL-DOX induced LV dysfunction with marked impairment in myocardium strain and synchronicity.

Fig. 2.

Doxorubicin (DOX) impaired cardiac strain and synchronicity and thereby left ventricular (LV) dysfunction. A: representative immunofluorescence images (×40) of the myocardium displaying α-smooth muscle actin (α-SMA) expression (red) and nuclei (blue) in the acute high-dose DOX (AH-DOX) and latent low-dose DOX (LL-DOX) groups compared with the saline control group. Data are representative from n = 2–3 mice/group. B: LV speckle tracking-based longitudinal B-mode echocardiographic images from the AH-DOX and LL-DOX groups compared with the saline control group. C: representative three-dimensional (3D) myocardial strain in AH-DOX and LL-DOX groups compared with the saline control group. D: LV segmental synchronicity images from saline control, AH-DOX, and LL-DOX groups showing impaired strain driving toward LV dysfunction because of DOX treatment. n = 8 mice/group (with quantitative data presented in Table 2).

Table 1.

Echocardiographic parameters in the AH-DOX and LL-DOX groups compared with the saline control group

Necropsy parameters	Saline Control Group	AH-DOX Group	LL-DOX Group
Sample size	6	6	7
Heart rate, beats/min	488.2 ± 4.8	474 ± 24.8	488 ± 51.1
End-diastolic dimension, mm	3.6 ± 0.24	3.8 ± 0.10	3.2 ± 0.14*
End-systolic dimension, mm	2.4 ± 0.20	2.9 ± 0.13	2.3 ± 0.14
Fractional shortening, %	33 ± 0.03	23.10 ± 2.39*	29.8 ± 1.87*
Global longitudinal strain	−15.0 ± 1.43	−13.6 ± 1.16*	−16.4 ± 2.14
Time to peak	57.8 ± 3.92	54.4 ± 3.06*	60.8 ± 4.96

Values are means ± SE. AH-DOX, acute high-dose doxorubicin; LL-DOX, latent low-dose doxorubicin. Data were analyzed using one-way ANOVA (Graph pad prism 7.0).

^*P < 0.05 vs. the saline control group.

DOX-induced irreversible dysregulation of LOXs and COXs in the LV.

LV functional and histological analysis revealed LV dysfunction with increased myocardium apoptosis and fibrosis with AH-DOX and LL-DOX treatment. Thus, we further added gravimetric analyses of the LV, spleen, and lung masses in the AH-DOX and LL-DOX groups. Both AH-DOX and LL-DOX groups demonstrated reduced LV mass normalized with body weight (P < 0.01) and spleen mass that was also normalized with body mass (P < 0.01) compared with the saline control group, indicating signs of splenocardiac cachexia (Table 2). LV cachexia was strongly supported by disorganization of the myocardium with myocyte atrophy in AH-DOX- and LL-DOX-treated mice, as supported by WGA staining data. AH-DOX and LL-DOX groups displayed reduced cardiomyocyte size compared with the saline control group (Fig. 3, A and B). Next, to determine integrative immunometabolic regulation, LOXs and COXs were determined since LOX- and COX-mediated muscle loss is a well-known mechanism in animal models (3, 5). In the present study, AH-DOX- and LL-DOX-treated mouse LVs displayed significant decreases in mRNA expression of ALOX5 (P < 0.001) and ALOX12 (P < 0.001) with no significant change in ALOX15 expression compared with saline control LVs (Fig. 3C). Furthermore, mRNA expression of COX [COX-1 (P < 0.01) and COX-2 (P < 0.5)] was decreased in the LV of AH-DOX and LL-DOX compared with saline control groups (Fig. 3D). These results indicate that DOX induced cardiac cachexia with a marked alteration of immune-responsive LOX and COX dysregulation, thereby suggestive of defective immunometabolism.

Table 2.

Gravimetric and necropsy parameters in the AH-DOX and LL-DOX groups compared with the saline control group

Parameters	Saline Control Group	AH-DOX Group	LL-DOX Group
Sample size	10	5	7
Body weight, g	25 ± 0.57	24 ± 0.67	19.2 ± 1.4*
Left ventricle, mg	82 ± 4	79 ± 4	60 ± 6*
Left ventricle/body weight, mg/g	3.3 ± 0.11	3.2 ± 0.12	2.3 ± 0.1*
Lung mass (wet), mg	172 ± 20	160 ± 15	146 ± 19
Lung mass (dry), mg	34 ± 2	37 ± 2	25 ± 1*
Spleen, mg	78 ± 2	68 ± 2	23 ± 2*
Spleen/body weight, mg/g	3.0 ± 0.38	2.7 ± 0.05	1.6 ± 0.29*

Values are means ± SE. AH-DOX, acute high-dose doxorubicin; LL-DOX, latent low-dose doxorubicin. Data were analyzed by using one-way ANOVA (Graph pad prism 7.0).

^*P < 0.05 vs. the saline control group.

Fig. 3.

Doxorubicin (DOX) induced cardiac myocardium atrophy and dysregulated genes encoding lipoxygenases (LOXs) and cyclooxygenases (COXs) in the left ventricle (LV). A: immunofluorescence images representing wheat germ agglutinin (WGA)-stained cardiac myocyte walls (green). LV cardiomyocyte nuclei were stained with Hoechst (blue; ×40) in saline control, acute high-dose DOX (AH-DOX), and latent low-dose DOX (LL-DOX) groups. B: representative bar graph of cardiomyocyte area in AH-DOX and LL-DOX groups compared with the saline control group (n = 5 mice/group). C: mRNA expression of LOXs [arachidonate 5-lipoxygenase (ALOX5), arachidonate 12-lipoxygenase (ALOX12) and arachidonate 15-lipoxygenase (ALOX15); middle] in the LV from saline control, AH-DOX, and LL-DOX groups. D: mRNA expression of COXs (COX-1 and COX-2) in the LV from saline control, AH-DOX, and LL-DOX groups. mRNA levels were normalized to hypoxanthine phosphoribosyltransferase 1. Values are means ± SE; n = 3–5 mice/group *P < 0.05 vs. the saline control group (analyzed by one-way ANOVA). E: immunoblot representing COX-1 expression in the LV of saline control, AH-DOX, and LL-DOX groups. F: densitometric analysis of COX-1 expression normalized to total protein. n = 5 mice/group. *P < 0.05 vs. the saline control group (analyzed by one-way ANOVA). ns, Not significant.

AH-DOX and LL-DOX induced splenic contraction and intoxicated CD169⁺ macrophages in the marginal zone.

The splenic leukocyte reservoir defines the cardiac healing and immune defense in infection (10, 26). Histological analysis of spleens from the saline control group showed a precise splenic organization with distinct identifiable regions of red and white pulp. Spleens from AH-DOX and LL-DOX-treated mice showed expanded red pulp with smaller encapsulated germinal center characteristics compared with saline control-treated spleens. DOX-mediated histological disorganization of the spleen was prominent in LL-DOX spleens (Fig. 4A). The spleen is a reservoir for leukocytes and propagates the immune response in injury; the murine marginal zone contains marginal metallophilic macrophages that are CD169⁺, a subpopulation of macrophages (F4/80⁺), and regulate an adaptive immune response. Immunofluorescence microscopy of F4/80⁺ (green) and CD169⁺ (red) macrophages in the spleen demonstrated depletion of CD169⁺ cells in the AH-DOX and LL-DOX groups (Fig. 4B). Furthermore, mRNA analysis of genes encoding COX-1, ALOX12, and ALOX15 showed that both AH-DOX and LL-DOX enhanced mRNA expression of COX-1 (P < 0.05, ~3.6-fold for the LL-DOX group), ALOX12 (P < 0.01, ~23-fold for the AH-DOX group and ~43-fold for the LL-DOX group), and ALOX15 (P < 0.01, ~20-fold for the AH-DOX group and ~15-fold for the LL-DOX group) compared with the saline control groups in the spleen, in contrast to their respective levels in the LV (Fig. 4C). These results indicate that DOX-induced splenocardiac toxicity is marked by depletion of CD169⁺ macrophages that are essential for the initiation of the immune response postinfection or injury.

Fig. 4.

Doxorubicin (DOX) induced splenic hypoplasia and contraction and compacted the germinal center. A: representative hematoxylin and eosin (H&E)-stained (left) spleen images at ×10 magnification accompanied with images at ×1.25 magnificaton of acute high-dose DOX (AH-DOX) and latent low-dose DOX (LL-DOX) groups compared with the saline control group. Scale bar = 100 μm. n = 5 mice/group. B: spleen immunofluorescence images representing F4/80⁺ (green) and CD169⁺ (red) in AH-DOX and LL-DOX groups compared with the saline control group. B: spleen mRNA expression of cyclooxygenase-1 (COX-1), arachidonate 12-lipoxygenase (ALOX12), and arachidonate 15-lipoxygenase (ALOX15) in saline control, AH-DOX, and LL-DOX groups. mRNA levels were normalized to hypoxanthine phosphoribosyltransferase 1. Values are expressed as means ± SE; n = 5 mice/group. *P < 0.05 vs. the saline control group (analyzed by one-way ANOVA).

DOX-induced cardiac toxicity dysregulated immune-responsive metabolic LMs.

Immune-responsive COX and LOX are essential to metabolize fatty acids in response to injury, infection, stress, and exercise with the biosynthesis of LMs (10, 14). DOX drastically reduced cardiac LOX levels. We further quantified levels of LMs in the LV using LC-MS/MS. The comprehensive lipidomic outcome revealed that the LOX-derived LMs 4-HDoHE, 16-HDoHE, and 12-HEPE were lowered in AH-DOX and LL-DOX groups compared with the saline control group (Fig. 5A). The overall pie chart distribution showed the presence of LOX-derived LMs (16-HoDHE: 16%, 4-HoDHE: 7%, and 12-HEPE: 12%) with minimal or no changes in 16-HoDHE and 4-HoDHE distribution in AH-DOX and LL-DOX groups, but 12-HEPE was lowered by 8% in the LL-DOX group (Fig. 5B). Of note, HoDHE and HEPE are LMs of essential fatty acids and precursors for D- and E-series resolvins, respectively. Therefore, DOX-induced pathology was impaired long-term host defense and there was an imbalance of LOX-mediated immune-responsive LMs in the LV. Analysis of COX-mediated LMs showed a decrease in levels of PGD₂ and PGE₂ in both AH-DOX and LL-DOX groups compared with the saline control group. However, the 6-keto-PG_2α level was lowered only in the LL-DOX group (Fig. 5C). The pie chart distribution of three LMs in LV samples of the saline control, AH-DOX, and LL-DOX groups did not demonstrate any difference (Fig. 5D). Together, these results suggest that DOX treatment caused an imbalance in immune-responsive LOX and COX expression and thereby impaired LM levels, indicating that it may be an early event of defective immunometabolism in DOX-induced cardiotoxicity.

Fig. 5.

Doxorubicin (DOX) impaired lipoxygenase (LOX)- and cyclooxygenase (COX)-mediated lipid mediators. A: bar graphs representing left ventricular (LV) LOX pathway metabolites 16-HDoHE, 4-HDoHE, and 12-HEPE in acute high-dose DOX (AH-DOX) and latent low-dose DOX (LL-DOX) groups compared with the saline control group. Quantification and values were ng/50 mg LV tissue. Values are means ± SE; n = 4 mice/group. *P < 0.05 vs. the saline control group (analyzed by one-way ANOVA). B: pie charts representing LOX-derived bioactive lipid mediators in the saline control, AH-DOX, and LL-DOX groups. C: bar graphs representing LV COX pathway prostaglandins (PGD₂, PGE₂, and 6-keto-PG_2α) in AH-DOX and LL-DOX groups compared with the saline control group. Quantification and values were ng/50 mg LV tissue. Values are expressed as mean ± SE; n = 4 mice/group. *P < 0.05 vs. the saline control group (analyzed by one-way ANOVA). D: pie charts representing COX-derived prostaglandin lipid species in the saline control, AH-DOX, and LL-DOX groups. ns, Not significant.

DOX-depressed CD169⁺ splenic macrophages with dysregulation of splenocardiac pro- and antiapoptotic gene expression.

Our histological examination revealed that DOX lowered CD169⁺ macrophages in the splenic marginal zone area. Therefore, to validate that DOX induced dysregulation in the spleen, we used a human translational dose (5 mg/kg) and a latent dose (7.5 mg/kg) in an acute setting (24 h). Quantitative results of flow cytometry confirmed that DOX reduced splenic CD169⁺ macrophages within 24 h. Both DOX doses (5 and 7.5 mg/kg) produced 3.7 ± 0.2% and 3.4 ± 0.1%, respectively, compared with saline controls, which displayed 5.8 ± 0.2% (Fig. 6, A and B). Since DOX-induced cardiotoxicity was observed by an increase in TUNEL-positive cells in the LV compared with saline-injected mice, that prompted us to determine the expression levels of programmed cell death markers in both the LV and spleen. The gene encoding the mRNA expression of the proapoptotic marker Bax was upregulated, and antiapoptotic Bcl2 expression was downregulated in both AH-DOX and LL-DOX groups in the LV and spleen compared with the saline control group (Fig. 6, C–F). Thus, quantitative analyses of CD169⁺ macrophages and the measurement of pro- and antiapoptotic gene expression supported DOX-induced splenocardiac toxicity.

Fig. 6.

Doxorubicin (DOX) intoxicated splenic macrophages and dysregulated splenocardiac pro- and antiapoptotic gene expression. A: representative FACS dot plot showing CD169⁺ macrophages in saline control and DOX (5 and 7.5 mg/kg dose)-treated groups. n = 4 mice/group. B: bar graph representing quantification of CD169⁺ macrophages in DOX (5 and 7.5 mg/kg dose)-treated groups compared with the saline control group. n = 4 mice. *P < 0.05 vs. the saline control group. C−E: genes encoding mRNA expression of Bax in the left ventricle (C), Bcl-2 in the left ventricle (D), and Bax in the spleen (E). F: spleen Bcl-2 (F) levels in the saline control, acute high-dose DOX (AH-DOX), and latent low-dose DOX (LL-DOX) groups. mRNA levels were normalized with hypoxanthine phosphoribosyltransferase 1. Values are expressed as means ± SE; n = 3–5 mice/group. *P < 0.05 vs. the saline control group (analyzed by one-way ANOVA).

DOX-induced cytokine and chemokine dysregulation in the LV and spleen.

Since DOX treatment altered immune-responsive LOXs and COXs in the LV and spleen, we further determined how DOX impacted cytokine and chemokine profiling. Both AH-DOX and LL-DOX repressed physiological levels of TNF-α (P < 0.05) mRNA expression in the spleen compared with the saline control group, indicative of a decreased residential leukocyte population and immunosuppression (Fig. 7A). In the LV, both AH-DOX and LL-DOX groups displayed a decrease in mRNA levels of mannose receptor C-type 1 (MRC-1; P < 001) compared with the saline control group, indicating intoxication of macrophages (Fig. 7B). Thus, the presented results suggest that DOX not only altered LOX and COX expression but also dysregulated cytokines and chemokines. Overall, the dysregulated physiological levels of cytokines and chemokines in the LV and spleen may be suggestive of integrative splenocardiac pathology.

Fig. 7.

Doxorubicin (DOX) suppressed physiological levels of TNF-α and the reparative marker mannose receptor C-type 1 (MRC-1) in the spleen and left ventricle (LV), respectively. A: mRNA expression of TNF-α in the spleen. B: MRC-1 in LVs from the saline control, acute high-dose DOX (AH-DOX), and latent low-dose DOX (LL-DOX) groups. mRNA levels were normalized with hypoxanthine phosphoribosyltransferase 1. Values are expressed as means ± SE; n = 5 mice/group. *P < 0.05 vs. the saline control group (analyzed by one-way ANOVA). C: the schematic design shows the novel mechanism of DOX-induced dysregulation of cyclooxygenase (COX)- and lipoxygenase (LOX)-derived bioactive lipid mediators with defective immunometabolism and splenocardiac cachexia. LMs, lipid mediators.

DISCUSSION

The chemotherapeutic agent DOX, from the class of anthracyclines, is commonly used in oncological practice. However, its clinical use is restricted due to the immediate or long-term risk of cardiotoxic properties because of myelosuppression in patients (2, 17, 29, 40). The present study was designed to investigate an integrative mechanism of DOX-induced cardiomyopathy with a major focus on LOX- and COX-mediated dysregulation and thereby defective immunometabolism. H-DOX (15 mg/kg) and LL-DOX (7.5 mg/kg) treatment 1) distinguished the survival and nonsurvival dose with irreversible systems-wide integrative splenocardiac pathology, 2) produced splenic hypoplasia with impaired immune-sensitive COX and LOX enzymes and thereby reduced bioactive LMs, and 3) intoxicated splenic macrophages and facilitated splenic contraction, thereby causing comprehensive splenocardiac pathology (Fig. 7C). DOX-induced cardiotoxicity is multifactorial and may involve various signaling mechanisms, including oxidative stress, mitochondrial dysfunction, apoptosis, iron homeostasis, and Ca²⁺ overloading (7, 21, 33, 35, 41). In the present report, we identified three novel DOX-induced pernicious actions that defined an integrative and comprehensive mechanism in the spleen and heart.

The first mechanism of action is DOX-induced irreversible dysregulation of COXs and LOXs, thereby causing defective immunometabolism. Our AH-DOX and LL-DOX experiments provided a quantitative measurement of immune-responsive lipid metabolites using LC-MS/MS in the myocardium. COXs (1, 2) and LOXs (5, 12, 15) are immune-sensitive metabolic enzymes required for the biosynthesis of autacoids using essential fatty acids that are enriched in the spleen. DOX-impaired arachidonic acid metabolism with reduced levels of PGD₂, PGE₂, and 6-keto-PG_2α and docosahexaenoic acid metabolism with lowered levels of 4-HDoHE and 16-HDoHE thus depressed levels of resolvin precursors in cardiac tissue. AH-DOX and LL-DOX also impacted eicosapentaenoic acid metabolite 12-HEPE, and this was indicative of comprehensive decreased immunometabolism. In cardiac injury, essential fatty acid-derived leukotriene B₄ is depleted in the spleen and enriched in the myocardium, which is essential for leukocyte recruitment, and, equally, lipoxin A₄ facilitates deceased cardiomyocyte clearance. Thus, the balance of inflammation-resolution mediators is essential for cardiac homeostasis (10).

The second DOX mechanism of action is that DOX intoxicated splenic marginal zone macrophages, thereby causing splenic contraction. In response to cardiac injury, splenic leukocytes activate the resolution process, as similarly reported in infection, where CD169⁺ macrophages control the pathogen to facilitate a host defense program (10, 26). Since splenic leukocytes have the unique capacity through the neural network to sense cardiac or stroke injury and/or infection, immediate deployment of monocytes for the biosynthesis of bioactive lipids at the site of injury (myocardium infarct in heart and brain infarct in stroke) occurs to facilitate the clearance and resolution phenomenon (10). DOX-mediated splenic mass contraction, germinal center compaction, and depletion of CD169⁺ macrophages were suggestive of DOX toxicity to impair the host defense system and immunosuppression. Both AH-DOX- and LL-DOX-injected mice dysregulated COXs and LOXs in the spleen and decreased CD169⁺ macrophages, supportive to nonresolving inflammation milieu.

The third DOX mechanism of action is the obvious consequence of immune cell population in splenocardiac toxicity, evident because of an imbalance of chemokines and cytokines. Activation of pyrogenic TNF-α is essential to leukocyte activation to facilitate cardinal signs of inflammation and host defense (10, 26). However, DOX suppressed physiological levels of cytokines, indicative of immunosuppression or impaired immune cell secretion. In contrast, DOX induced irreversible suppression of the reparative marker MRC-1/CD206, which is suggestive of advanced myelosuppression because of defective immunometabolism. Thus, DOX suppressed the immune defense capacity of splenic leukocytes, which is essential to facilitate the clearance of preneoplastic and mature cancer cells (6). DOX mediated the decreased reparative marker MRC-1/CD206 in the heart and physiological downregulation of TNF-α in spleens from both AH-DOX and LL-DOX groups, indicative of an adverse impact on immune cell health.

In summary, our study introduces a novel and integrative approach to define the splenocardiac impact of an oncological drug using DOX as an example in a noncancer model. Here, we identified and outlined the translational mechanism of DOX-induced splenocardiac cachexia with defective immunometabolism in cardiac toxicity. Based on our findings, future strategies are warranted to preserve immune and cardiac health using cancer models.

GRANTS

This work was supported in part by National Heart, Lung, and Blood Institute Grant HL-132989 (to G. V. Halade).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

G.V.H. conceived and designed research; J.K.J., G.W.W., V.K., M.A.S., and G.V.H. performed experiments; J.K.J., G.W.W., V.K., and G.V.H. analyzed data; V.K., N.Y., and G.V.H. interpreted results of experiments; J.K.J., G.W.W., V.K., and G.V.H. prepared figures; J.K.J., V.K., N.Y., and G.V.H. drafted manuscript; N.Y. and G.V.H. edited and revised manuscript; G.V.H. approved final version of manuscript.