The authors have read the journal’s policy and have the following conflicts: Authors Dr. Vacchi-Suzzi, Dr. Moggs, Dr. Pognan, Dr. Grenet, Dr. Couttet, Dr. Bongiovanni, Dr. Letzkus and Dr. Staedtler are employed by Novartis Institutes for Biomedical Research. Dr. Bauer is employed by Actelion Pharmaceuticals Ltd. Dr. Berridge, Dr. Lyon and Dr. Vidgeon-Hart are employed by GlaxoSmithKline. Dr. Hamadeh is employed by Amgen Inc. This does not alter the authors' adherence to all the PLoS ONE policies on sharing data and materials.
Conceived and designed the experiments: HH JL RP PC CVS. Performed the experiments: BRB ML MPV-H CVS. Analyzed the data: YB KG HH CVS. Contributed reagents/materials/analysis tools: SB FS. Wrote the paper: CVS KG PC. Supervised the study and experimental design, helped interpretation of the data: CVS PC JM OG FP YB KG HH JL RP NH FS. Reviewed and commented on the manuscript: CVS YB BRB SB KG HH ML JL JM RP FP FS MPV-H OG PC.
Anti-cancer therapy based on anthracyclines (DNA intercalating Topoisomerase II inhibitors) is limited by adverse effects of these compounds on the cardiovascular system, ultimately causing heart failure. Despite extensive investigations into the effects of doxorubicin on the cardiovascular system, the molecular mechanisms of toxicity remain largely unknown. MicroRNAs are endogenously transcribed non-coding 22 nucleotide long RNAs that regulate gene expression by decreasing mRNA stability and translation and play key roles in cardiac physiology and pathologies. Increasing doses of doxorubicin, but not etoposide (a Topoisomerase II inhibitor devoid of cardiovascular toxicity), specifically induced the up-regulation of miR-208b, miR-216b, miR-215, miR-34c and miR-367 in rat hearts. Furthermore, the lowest dosing regime (1 mg/kg/week for 2 weeks) led to a detectable increase of miR-216b in the absence of histopathological findings or alteration of classical cardiac stress biomarkers.
Doxorubicin (DOX), a potent drug used in cancer chemotherapy, intercalates with DNA and stabilizes a ternary complex with Topoisomerase II (Top2) thus preventing replication of DNA and subsequent cell proliferation
Despite its beneficial therapeutic effects, acute administration of high DOX doses causes severe kidney damage, whilst toxic cardiomyopathy is observed when DOX is chronically administered for several weeks at doses devoid of severe nephrotoxicity
Administration of DOX ranging from 1 to 3 mg/kg/week to rats during several weeks recapitulates the cardiac symptoms observed in humans
MicroRNAs are a class of genomically encoded non-translated short RNAs, about 22 nucleotides long, discovered in plants in 1993
Recent studies investigating DOX-induced changes in gene expression at the mRNA level in rat cardiac tissue have successfully identified genomic biomarkers whose molecular functions are consistent with proposed toxicity mechanisms
In the present study, we report that the chronic myocardial toxicity induced by DOX in rats was associated with the modulation of microRNAs, and some of these can be phenotypically anchored to histopathology findings. Alteration of miR-216b could be detected earlier than overt myocardial damage. Furthermore, by integrating DOX-inducible microRNAs with mRNA profiles generated from the same cardiac tissue samples, the putative target genes of the DOX-induced microRNA signature were highlighted, including a number of genes that have not previously been described in DOX-induced cardiomyopathy. Finally, evidence for the direct targeting of Sipa1 mRNAs by miR-34c was provided for the first time, and this microRNA-mRNA interaction could potentially play a role in the molecular response to DOX in the rodent heart.
Male rats were treated with DOX according to the scheme reported in
(A) Six adult male rats were injected with the indicated doses of vehicle, doxorubicin (DOX), dexrazoxane (DZR), etoposide (EPS) or a combination of DOX and DZR for 2, 4 or 6 weeks. Cardiac tissue was excised and deep frozen for gene expression and microRNA profiling experiments. A representative micrograph of a toluidine blue stained myocardial section of a control (B) and of a DOX treated animal (C). Black arrows indicate sarcoplasmic micro- and macro- vacuolation of cardiomyocytes.
The administration of DOX at 3 mg/kg/week induced time-dependent changes in the expression of mRNAs that have previously been associated with molecular responses underlying cardiac pathologies
Expression fold change relative to vehicle were represented for DOX 3 mg/kg/week at 2, 4 and 6 weeks time point (n = 6) for (A) Ankrd/Carp, (B) Nppb, (C) Myh7 and (D) Myh6. For each time point and each probe set, vehicle values were averaged and normalized to 1. The same correction was applied to the DOX treated values. Affymetrix probe-set number is indicated in brackets. Error bars represent standard deviation. T-test was performed for vehicle- vs. DOX-treated at each time point. *P<0.05, **P<0.01, ***P<0.005, NS = Non-Significant. (No t-test for #, as n = 2).
Expression profiles of 518 rodent microRNAs were generated by using a low density array qPCR platform (TLDAs). Approximately 370 microRNAs gave a measurable signal below the arbitrary cut-off set at 35 qPCR cycle times (Ct), with no major distinction between controls and treated samples. The administration of 4 weekly doses of DOX 3 mg/kg/week led to increased levels of 17 microRNAs and decrease levels of 8 microRNAs (p value below 0.05) (
2 weeks | 4 weeks | |||
DOX 3 mg/kg/week | FC | P value | FC | P value |
|
4.58 | 1.97E−01 | 74.69 | |
mmu-miR-215 |
1.11 | 9.41E−01 | 51.42 | |
|
42.58 | 24.17 | ||
|
2.22 | 5.69E−01 | 21.5 | |
|
1.79 | 5.16E−01 | 17.71 | |
|
1.74 | 1.07E−01 | 12.69 | |
mmu-miR-667 | 1.49 | 3.08E−01 | 7.19 | |
mmu-miR-298 | −1.44 | 1.83E−01 | 3.67 | |
|
1.6 | 8.87E−02 | 3.42 | |
mmu-miR-208b |
1.09 | 7.94E−01 | 3.23 |
|
mmu-miR-21* | 1.16 | 7.36E−01 | 3.21 | |
|
1.55 | 1.79E−01 | 2.74 | |
mmu-miR-709 |
1.04 | 8.38E−01 | 2.51 |
|
mmu-miR-708 |
1.11 | 7.08E−01 | 2.34 | |
mmu-miR-31 | −1.02 | 9.68E−01 | 2.32 | |
mmu-miR-34c |
1.02 | 9.55E−01 | 2.25 | |
rno-miR-29b-2* | −1.1 | 6.84E−01 | 2.01 | |
mmu-miR-218-2* |
1.03 | 9.79E−01 | −8.61 |
|
|
−3.21 | 3.29E−01 | −6.14 | |
rno-let-7e* | 1.49 | 6.00E−01 | −3.95 | |
|
−2.86 | 2.30E−01 | −3.76 | |
mmu-miR-335-3p | 1.43 | 1.33E−01 | −2.52 | |
mmu-miR-384-3p | 1.07 | 7.52E−01 | −2.38 | |
mmu-miR-221 | 1.37 | 1.15E−01 | −2.08 | |
rno-miR-1* | −1.16 | 8.05E−01 | −2.01 |
Variation of cardiac microRNA levels versus vehicle are reported for animals treated with DOX 3 mg/kg/week for 2 and 4 weeks. Values were calculated via the relative quantification (ΔΔCt) method by using the mammalian U6 snRNA as a normalizer. MicroRNAs showing same trend at 2 and 4 weeks are italicized. # indicates microRNAs selected for further analysis. Significant P values (<0.05) are in bold. FC = fold change.
A subset of ten microRNAs whose expression was altered by chronic DOX treatment was selected based on their lowest p values (
Relative quantification of (A) miR-208b, (B) miR-215, (C) miR-216b, (D) miR-367 and (E) miR-34c in DOX, DOX + DZR, EPS groups, normalized versus vehicle treated animals. Expression levels were measured by single assay qPCR (n = 3, except #, n = 2). DOX: Doxorubicin, DZR: dexrazoxane, EPS: etoposide; numbers indicate the weekly dose of each compound in mg/kg/week. Empty spaces represent non-sampled animals. The vehicle treated is the first column of each time-point. The animals used in this experiment were distinct from the ones represented in
Following two weeks treatment with DOX alone, a slight up-regulation of miR-208b, mir-216b and miR-367 was observed in the heart of most animals. Following 4 weeks of treatment with DOX alone, the level of all five microRNA candidates were significantly increased and dose dependency was observed for miR-208b, miR-215, miR-216b and miR-367. Following 6 weeks of treatment with DOX alone, the levels of miR-208b, miR-215, miR-216b and miR-367 were significantly increased. Dose dependency was difficult to assess at the 6 week time point since only 2 animals survived in the high dose group.
Following treatment with DZR alone for 2 and 6 weeks, none of the five candidate microRNAs was affected. After treatment with the combination of DOX at 2 mg/kg/week and DZR at 50 mg/kg/week, the level of miR-208b, miR-216b and miR-367 was either slightly increased or not affected. For all five candidate microRNAs, the level was lower when compared to the group of animals treated for 2 weeks with DOX alone at 2 mg/kg/week. Following 6 weeks of treatment with the same DOX/DZR combination, the level of miR-208b, miR-216b and miR-367 was increased. However, this level was lower for miR-208b and miR-216b when compared to the group of animals treated 6 weeks with DOX alone at 2 mg/kg/week and higher for miR-367.
Following 2 and 6 weeks of treatment with EPS, the expression level of the 5 microRNA candidates was minimally affected and considered as non statistically significant since no dose and time dependency was observed (although these changes might still be biologically relevant). Together, these data suggested that the variation of these five microRNAs was DOX-specific.
In summary, chronic treatment with DOX at 3 mg/kg/week for 2 weeks led to increases in miR-208b, miR-216b and miR-367, with fold change intensities comparable to those shown by genomic cardiomyopathy indicators (Ankrd/Carp, Nppb, Myh7 and Myh6) (
We then compared the variation of DOX-specific microRNA perturbations with the severity of cardiac tissue lesions induced by drug (cytoplasm vacuolation) in order to evaluate their sensitivity. Increasing doses of DOX led to higher cumulative vacuolation grading with time (
Blue bars show cumulative vacuolation grade. X axis shows the DOX regimen in mg/kg/week received by the animals at 2 weeks. Y axes report cumulative histopathological scores and microRNA fold change vs. untreated cardiac tissues (normalized at value 1). Path grading = cumulative vacuolation score. FC = fold change. SEM = standard error on the mean.
Out of the 7804 mRNA targets for DOX-inducible microRNAs predicted by Microcosm, 132 (1.7%) mRNAs showed the expected profile of anti-correlated expression relative to its putative targeting microRNA (>1.5 fold change at DOX 3 mg/kg/week for 4 weeks
(A) Sipa1 mRNA raw expression values decreased in the heart of rats treated with DOX. #, n = 2. (B) DOX treatment for 24 h caused a decrease of Spa1 mRNA and an increase of miR-34c in cardiac myoblast cells (H9c2). (C) H9c2 endogenous Sipa1 mRNA was decreased by transfection of miR-34c mimic, and increased using a miR-34c hairpin inhibitor (HI). Transfection with miR-34c mimic and inhibitor respectively exacerbated and rescued Sipa1 mRNA levels in H9c2 treated with DOX 0.1 and 1 µM overnight in comparison to negative controls. (D) Alignment of mammalian miR-34 family. Capital letters indicate mismatch in the sequence. Sipa1 3′-UTR wt and MUT construct: 12 nt surrounding the predicted seed are shown. HEK 293 cells were co-transfected with pmiR-GLO-Sipa1 and the indicated miRNA mimic or a
We have also used the rat myoblast H9c2 cell line to assess the DOX-induced autophagy process through the regulation of Ambra1 mRNA (
(A) Fold change of Ambra1 probe set in rat heart tissue treated with DOX. Fold change and statistical significance were assessed
In the present study 25 microRNAs were found to be dysregulated by pharmacological intervention with DOX in the rat heart. A subset of microRNA candidates were further investigated and associated to their putative targets.
The chronic weekly intravenous administration of DOX was designed to trigger cardiomyopathy symptoms within a 6 week time-frame allowing us to identify mRNAs and microRNAs that are altered in the early stages of the drug-induced cardiac remodeling. Cardiac toxicity at a cumulative dose of 12 mg/kg (3 mg/kg/week for 4 weeks) was confirmed by histopathological lesions and altered expression of genomic indicators associated to contraction performance (Myh6, Myh7) and heart failure (Ankrd1/CARP, Nppb) (
We profiled hundreds of microRNAs via TLDAs in the heart of rats that showed clear signs of toxicity (DOX 3 mg/kg for 4 weeks), and found that 25 microRNAs were dysregulated with p values below 0.05. Seven out of 25 microRNAs (miR-367, miR-216b, miR-383, miR-692, miR-135b, miR-145*, miR-877, miR-434-5p and miR-337-5p), were already modulated after 2 weeks of DOX at 3 mg/kg/week (
It is noteworthy that miR-21 and miR-146a, previously linked to DOX-induced apoptosis in cardiomyocytes
In our study miR-215, part of the miR-192/miR-215 cluster, and miR-34c, involved in DNA damage mediated proliferation arrest
Interestingly miR-215, up regulated by DOX in this study, was also linked to p53 mediated cell cycle arrest
MicroRNA-208b, encoded from the intron 28 of rat Myh7, is associated to maintenance of myocardial performance together with 2 other myomirs, i.e. miR-208a/miR-499, which play a pivotal role in the myosin balance
Of particular interest is the predictive potential of microRNAs as early tissue indicators of drug-induced cardiac lesions. The lowest dose of DOX tested in our study (1 mg/kg/week for 2 weeks) did not cause detectable tissue vacuolation (
In order to gain further insight into the cellular processes that may be modulated upon DOX-induced cardiac toxicity, high confidence target mRNAs of DOX-dysregulated microRNAs were identified by using the Microcosm prediction tool (
In conclusion, 25 microRNAs implicated in DOX-induced cardiac toxicity
Our study also provides the basis for understanding the global role of microRNAs in DOX-induced cardiac remodeling and toxicity, and provides evidence the direct miR-34c/Sipa1 functional interaction for the first time.
This in vivo study was conducted in compliance with the Animal Welfare Act, and the Office of Laboratory Animal Welfare, after review and approval by the Covance site (Vienna, Virginia) Institutional Animal Care and Use Committee. Doxorubicin (Adriamycin, Lot No. 86G23FY, CAS# 23214-92-8), a generous gift from Adria Laboratories Inc. (Columbus, OH, USA), etoposide (ETOPOPHOS® Lot No. 5E04155, CAS# 117091-64-2) was purchased from Bristol-Myers Squibb (Princeton, NJ, USA) and dexrazoxane (Zinecard® Lot No. ADR074B, CAS# 24584-09-6) purchased from Pharmacia/Pfizer (Kalamazoo, MI, USA) were administered once weekly as follows: 11 week old male Crl:CD(SD) rats received DOX or EPS via intravenous injection once a week for 2, 4, or 6 weeks based on
Total RNA was obtained from 50–100 mg frozen cardiac tissue homogenized in Trizol, according to manufacturer instructions (Invitrogen, Life Technologies, Carlsbad, CA). Long and short RNA fractions were separated via affinity resin column during clean up, according to the manufacturer instructions (miRNeasy Mini Kit, Qiagen, MD). Short RNA was quantified by absorbance at 260 nm using a Nanodrop (Thermo Scientific, Wilmington, DE), and the quality and integrity were determined using Small RNA chips (Agilent Technologies, Santa Clara, Ca) and stored at −80°C until analysis.
GeneChip experiment was conducted in the BMD Genechip laboratories (Novartis, BS) on Rat Genome Rat230 2.0 Array (Affymetrix, Inc.). Target preparation was performed with a starting amount of approximately 1 µg of total RNA unless otherwise specified using the Affymetrix GeneChip HT One-Cycle Target Labelling and control reagent according to manufacturer’s instruction (Affymetrix, Inc.). cRNA size distribution (before and after fragmentation) was confirmed by agarose-gel electrophoresis. An amount of biotinylated cRNA of approximately 10 µg for each sample was hybridized for approximately 16 hours at 45°C on an array. The array was washed and stained on Affymetrix Fluidics Workstation 450 and scanned on Affymetrix Scanner 3000 according to manufacturer’s technical manual. The scanned image was converted into numerical values of the signal intensity (Signal) and into categorical expression level measurement (Absolute Call) using the Affymetrix MAS 5.0 software. The software scaled the average intensity of each chip to a target intensity of 150. The microarray data are available in ArrayExpress under accession number E-MTAB-1168.
Using the Taqman Low Density Array (TLDA) technology (Applied Biosystems, Life Technologies, Carlsbad, CA), about 500 rodent microRNAs were profiled without any pre-amplification step, using 700 ng of small RNA according to the manufacturer instructions (Applied Biosystems, Life Technologies, Carlsbad, CA). Sufficient RNA amount for the TLDA cards was obtained for 3 out of 6 animals receiving DOX 3 mg/kg/week for 4 weeks, and for 6 out of 6 receiving DOX 3 mg/kg/week for 2 weeks. An arbitrary cut-off at 35 Ct was set in order to exclude microRNAs out of the linear range of detection. Mammalian snRNA U6 was used as housekeeping gene for its expression was not affected by DOX treatment. In order to confirm the results of the TLDA, a subset of microRNAs were quantified using single Taqman assays in the 3 remaining animals treated with DOX 3 mg/kg/week for 4 weeks. All procedures were carried out according to manufacturer instructions (Applied Biosystems, Life Technologies, Carlsbad, CA) using a 7900HT Fast Real-Time PCR System.
MicroRNA data obtained from the TLDAs were quality checked and analyzed using Statminer plugin (Integromics, Madrid, Spain) for TIBCO Spotfire (TIBCO Software, Palo Alto, CA). Average fold changes and standard deviations were calculated and plotted using GraphPad Prism 5.00 (GraphPad Software, San Diego California USA). Lists of putative targets for each DOX-regulated microRNA were obtained from Microcosm
HEK 293 (CRL-1573, ATCC) and rat myoblastic cells H9c2 (CRL-1446, ATCC) were cultured in MEM and DMEM respectively, supplemented with 10%v/v heat-inactivated fetal bovine serum (Invitrogen). DOX was purchased from Sigma, diluted in DMSO to a stock concentration of 10 mM and stored at −20°C. DOX treatment of H9c2 cells (n = 3) was independently repeated 3 times (total of 9 replicates). DOX was added cells transfected with microRNA reagents after 4 hours, buy replacing transfection medium with fresh growth medium supplemented with DOX at the indicated concentrations.
Two-hundred nucleotides surrounding the microRNA predicted seed site in the 3′-UTRs of either murine Tnni3k or Sipa1 were synthesized (GeneArt, Invitrogen, Carlsbad, CA) and cloned into a pmiR-GLO vector (Promega, Madison, WI) downstream of the luciferase gene via a XbaI-SacI restriction site to generate the pmiR-GLO-Sipa1 wt expression vector. The pmiR-GLO expression vector contains both the luciferase and the Renilla genes under the control of different promoters. Complete insert sequences are listed in
Sipa1 mutagenesis Primer1∶5′-
HEK 293 cells were plated in 24 wells plates 24 hours prior to co-transfection in antibiotic-free medium. Cells (2×105 per well) were co-transfected with the reporter constructs (20 ng/well) and synthetic mmu-miR-34a, b or c (40 or 80 nM) mimics (Dharmacon, Lafayette, CO) in serum-free medium, using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). A
(TIF)
(TIF)
(XLS)
(XLS)
(XLS)
(XLS)
(DOC)
(DOC)
The authors thank Dr. Raegan O’Lone for the excellent coordination between HESI partners. The authors would like to thank the HESI Application of Genomics to Mechanism-Based Risk Assessment Technical Committee for design and conduct of the in life study, providing samples for the described analysis.