Overexpression of multidrug resistance‐associated protein 1 protects against cardiotoxicity by augmenting the doxorubicin efflux from cardiomyocytes

Doxorubicin is a commonly used anti‐cancer drug used in treating a variety of malignancies. However, a major adverse effect is cardiotoxicity, which is dose dependent and can be either acute or chronic. Doxorubicin causes injury by DNA damage, the formation of free reactive oxygen radicals and induction of apoptosis. Our aim is to induce expression of the multidrug resistance‐associated protein 1 (MRP1) in cardiomyocytes derived from human iPS cells (hiPSC‐CM), to determine whether this will allow cells to effectively remove doxorubicin and confer cardioprotection. We generated a lentivirus vector encoding MRP1 (LV.MRP1) and validated its function in HEK293T cells and stem cell‐derived cardiomyocytes (hiPSC‐CM) by quantitative PCR and western blot analysis. The activity of the overexpressed MRP1 was also tested, by quantifying the amount of fluorescent dye exported from the cell by the transporter. We demonstrated reduced dye sequestration in cells overexpressing MRP1. Finally, we demonstrated that hiPSC‐CM transduced with LV.MRP1 were protected against doxorubicin injury. In conclusion, we have shown that we can successfully overexpress MRP1 protein in hiPSC‐CM, with functional transporter activity leading to protection against doxorubicin‐induced toxicity.


| BACKGROUND
Anthracyclines, a highly efficacious class of anti-cancer drugs, are used in the treatment of range of malignancies.However, one significant side effect of these drugs is dose-dependent cardiotoxicity, resulting from progressive and irreversible loss of cardiomyocytes, which are unable to regenerate. 1 Conventional neurohormonal therapies are required to manage symptoms, but ultimately do not ameliorate declining cardiac function, nor address the underlying cause of cardiotoxicity.
Gene therapy offers the prospect of confering cardioprotection during chemotherapy by rendering the heart resistant to anthracycline toxicity after one treatment.Successful protection relies on efficient vector delivery to the heart, with robust expression of the cardioprotective gene during the course of injury.
We explored the use of a drug transporter known as multidrug resistance-associated protein 1 (MRP1), which is known to confer drug resistance by directing removal of doxorubicin from cells. 2 Here, we demonstrated that overexpression of MRP1 using a lentiviral vector (LV.MRP1) resulted in augmented efflux activity in human induced-pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs).
Subsequently, intracellular levels of doxorubicin were reduced in MRP1 overexpressing hiPSC-CM, which led to increased survival.This study provides proof of concept of cardioprotective gene therapy which can protect against the cardiotoxic effects of anthracyclines.

| Maintenance of hiPS cells and differentiation into cardiomyocytes
Two different hiPSC lines were used in this study: the WTCWT iPSC line obtained from Professor Bruce Conklin (The J. David Gladstone Institutes) and the SCVI 8 iPSC line from Professor Joseph Wu (Stanford Cardiovascular Institute).4][5][6] Both lines were maintained on Matrigel (no.354277, Corning, New York, USA) coated 60 mm cell culture dishes using the mTeSR™ Plus kit (no.05825, STEMCELL Technologies, Vancouver, Canada).Upon confluence, cells were passaged as colonies using gentle cell dissociation reagent (no.07174, STEMCELL Technologies, Vancouver, Canada) every 6-7 days.
For cardiomyocyte differentiation, WTCWT cells were maintained and differentiated as previously described. 7,8For the SCVI 8 line, cells were dissociated from confluent dishes on D2 (2 days prior to differentiation start on D0) using TrypLE™ Express Enzyme (ThermoFisher Scientific, 12604-021, Massachusetts, USA), then seeded into Matrigel-coated 12-well plates at 700,000 cells/well using mTeSR™ Plus supplemented with Y-27632 (no.72304, STEM-CELL Technologies, Vancouver, Canada).On D1, a media change was performed to remove the Y-27632.On D0, the differentiation was then commenced using the STEMdiff Cardiomyocyte Differentiation Kit according to manufacturer's instructions until the point of maintenance in STEMdiff cardiomyocyte maintenance medium.On D11, beating cells were subjected to 2 days of metabolic selection using lactate medium composed of glucose-free DMEM (no.A14430-01, ThermoFisher Scientific, Massachusetts, USA) supplemented with L-(+)-lactic acid (no.L1750-10G, Sigma-Aldrich, St Louis, MO, USA) to final concentration of 4 mM and bovine serum albumin (no.A9418-10G, Sigma, St Louis, MO, USA) to 15 μM.On D13, cells were then returned to cardiomyocyte maintenance medium and maintained until differentiation completion on D15.

| Assessment of hiPSC-CM purity
Cells were dissociated using TrypLE Express, then washed with with Dulbecco's phosphate-buffered saline without calcium and magnesium (no.12001-664, Lonza, Basel, Switzerland).Cells were then stained using the Zombie NIR Fixable Viability Kit (no.423105, BioLegend, San Diego, CA, USA).After washing, cells were fixed in 4% PFA (w/v) for 30 min, then washed and further stained with BV421-conjugated mouse anti-cTnT antibody (no.565618, BD Biosciences, San Diego, CA, USA) for 2 h.Upon further washing, cells were analyzed on the FACSCanto II Cell Analyzer or LSRFortessa and data recorded using FACSDiva Software (BD Biosciences, Franklin Lakes, NJ, USA).Analysis was subsequently performed using FlowJo (Ashland, OR, USA) version 10.Only differentiation batches with >85% cTnT+ were used for subsequent experiments.

| Molecular cloning and lentiviral production
The pcDNA3.1(À)MRP1kplasmid was a gift from Professor Susan Cole (Queen's Cancer Research Institute).The MRP1 cDNA was cloned into the lentiviral vector plasmid pRRLsin18.cPPT.CMV.GFP.Wpre (pPPT.CMV.GFP, Inder Verma, The Salk Institute for Biological Studies, California, USA) after removal of green fluorescent protein (GFP).The new vector was named ppt.MRP1.A control vector was also generated, by including the LacZ coding sequence in place of the ghrelin transgene (ppt.LacZ).
Lentiviral vectors were produced by calcium phosphate transfection into HEK293T cells.Vector containing supernatant was collected at 48 and 72 h after transfection, then filtered and concentrated by ultrafiltration (100,000 MWCO PES, no.VS2042, Sartorius, Göttingen, Germany).Concentrated virus (LV.LacZ and LV.MRP1) was aliquoted and stored at À80 C. Transduction titer was assigned on concentrated supernatant by assessing transgene expression in HEK293T cells using a limiting dilution assay in the presence of polybrene 8 μg/mL (no.H9268-10G, Sigma-Aldrich, St Louis, MO, USA) 4 days after transduction.For transduction experiments, concentrated vector stock was used at the indicated multiplicity of transduction in the presence of polybrene 8 μg/mL.Vector was applied overnight and media changed the following morning.Cells were then harvested for various assays on day 4 post transduction.

| Amplification of MRP1 and GAPDH transcript by RT-PCR
RNA was extracted from transduced HEK293T and hiPSC-CM cells using the ISOLATE II RNA Micro Kit (no.BIO-52075, Bioline, Tennessee, USA) then quantified using a NanoDrop™ 2000 Spectrophotometer (no.ND2000, Thermo Fisher Scientific, Massachusetts, USA).For cDNA synthesis, Random Primers (500 ng/μL) were added to the extracted RNA (1 μg), followed by incubation at 70 C for 5 min.
The membrane was then blocked for 1 h in 5% skim milk, followed by overnight incubation with the primary mouse anti-MRP1 (no.ab24102, abcam, Massachusetts, USA) or rabbit anti-β-actin (no.ab8227, abcam, Massachusetts, USA) antibodies.The membrane was subsequently washed for 3 Â 10 min with PBST, followed by incubation with the rabbit anti-mouse (no.A9044-2ML, Sigma-Aldrich, St Louis, MO, USA) or goat anti-rabbit (no.P0448, Agilent Technologies, California, USA) secondary antibodies for 1 h.The membrane was then washed for 3 Â 10 min with PBST then imaged via chemiluminescence using the SuperSignal™ West Femto Maximum Sensitivity Substrate (no.34095, Thermo Fisher Scientific, Massachusetts, USA).
Transduced hiPSC-CMs in 24-well plates exposed to doxorubicin were also imaged and analyzed by flow cytometry as outlined above to evaluate red auto-fluorescence 48 h post injury.

| Doxorubicin injury of hiPSC-CMs
To investigate the effect of MRP1 overexpression, LV.LacZ or LV.
MRP1 transduced cells in 96-well plates were injured at D4 post-transduction with a dose titration of doxorubicin.After 48 h, cell viability was assessed using the CellTiter 96 ® non-radioactive cell proliferation assay (no.G4000, Promega, Wisconsin, USA), hereafter abbreviated to "MTT assay", according to manufacturer's instructions.
Cell viability was expressed as a percentage of the un-injured control (0 μM doxorubicin) within each treatment group.

| RESULTS
To test the function of the LV.MRP1 vector, overexpression of MRP1 was initially evaluated in HEK293T cells.Transduced cells showed significantly increased MRP1 mRNA expression 305.3 ± 16.22 fold higher than the non-transduced (NT) control (Figure 1A).This experiment was repeated in hiPSC-CM, which were differentiated and validated for cTnT+ purity of >85% prior to use for subsequent experiments (Figure S1).LV.MRP1 transduced hiPSC-CMs showed a 330.3 ± 83.99 fold increase of MRP1 transcript over NT (Figure 1B).Western blot analysis confirmed that MRP1 protein was produced only in the transduced hiPSC-CMs, with no signal observed in the NT or LV.LacZ control groups (Figure 1C).
The MRP1 protein was also shown to be functional, as LV.MRP1 transduced hiPSC-CMs were able to export fluorescence CDFA dye more efficiently than the NT control group (Figure 2A-C).As dye concentration increased, accumulation of CDFA increased significantly at 75 and 100 μM CDFA, while fluorescence was significant between groups at 100 μM CDFA (Figure 2B,C).In comparison, MRP1 expressing cells did not show increased accumulation of dye even up to 100 μM concentration (Figure 2C).
After proving the function of the recombinant MRP1, we then sought to determine whether this level of expression would allow cardioprotective efflux of doxorubicin from hiPSC-CMs.The natural red fluorescence of doxorubicin 9 allowed the direct visualization of drug accumulation in injured cells.While the overall number of doxorubicin positive cells did not change in both groups, there was a significant decrease in red fluorescence detected from LV.MRP1 transduced cells (Figure 3A).We observed that high levels of MRP1 overexpression were maintained in LV.MRP1 transduced hiPSC-CMs (254-711 fold increase over LV.LacZ 0 μM doxorubicin control) even after treatment with doxorubicin (Figure 3B).This was sufficient to augment efflux activity, leading to reduced red autofluorescence in the LV.MRP1 transduced cells compared with the LV.LacZ control group after 48 h of exposure to 10 μM doxorubicin (Figure 3C).
Overexpression of MRP1 also resulted in cardioprotection and amelioration of cell death in the LV.MRP1 cells, with a significant increase in survival at nearly every concentration (25-75 μM) of doxorubicin tested (Figure 3C).

| DISCUSSION
MRP1 is ubiquitously expressed in various organs in the body, including the heart. 10Under physiological conditions, MRP1 regulates redox balance within cells to protect against oxidative stress, by effluxing glutathione, glutathione disulfide and related conjugates. 11When the heart is exposed to doxorubicin, an acute response causes upregulation of MRP1 within 24 h of injury. 12However, this is counterbalanced by doxorubicin-mediated adduction of MRP1 with the reactive lipid peroxidation product 4-hydroxy-2-nonenal by 48-72 h post injury, which reduces efflux activity. 12Loss of MRP1 has also been shown to increase sensitivity of cardiac cells to doxorubicin mediated cardiotoxicity. 13 this study, we show the cardioprotective effect of constitutive study using MRP1 shows augmented efflux of doxorubicin as a mechanism for cardioprotection.However, MRP1 may be considered a difficult therapeutic target, as it does not efflux doxorubicin specifically.
MRP1 is also able to efflux a range of drugs and glutathione conjugates, which may in some cases prove detrimental in patients who are receiving concurrent alternative heart therapies. 15gardless of the therapeutic target, it is important to ensure that delivery is specific to the heart, to avoid inadvertent off-target gene tranfer to cancer cells, which would render them resistant to chemotherapy.Lentiviral vectors are able to package large genes such as MRP1, but they may not be the vector of choice for cardiac gene therapy, owing to limited in vivo transduction efficiency and risk of oncogenesis. 16Vectors with higher safety profile such as the adenoassociated virus may offer a way to to efficiently delivery transgenes to target cell populations. 17ile this study validates an in vitro platform for testing cardioprotective gene therapy, it does not allow further evaluation of off-target effects in non-cardiac organs.As such, the development of subsequent therapies would need to be tested in vivo to ensure safety as well as clinical benefit.It should also be noted that this study evaluated the cardioprotective effect on viability and not cell death and apoptosis.This is a significant limitation that would need to be addressed in subsequent studies, as apoptosis is the pathway most commonly associated with doxorubicin-mediated cell death in cardiomyocytes.

| CONCLUSION
Overexpression of MRP1 in hiPSC-CMs was cardioprotective against doxorubicin-induced damage.This provides proof of concept for the use of gene therapy as a protective strategy against anthracyclineinduced cardiotoxicity.

F I G U R E 1
Functional validation of lentivirus-mediated overexpression of multidrug resistance-associated protein 1 (MRP1).Lentivirus vector encoding MRP1 (LV.MRP1) was used to transduce (A) HEK293T cells (n = 3, multiplicity of infection (MOI) 10) and (B) cardiomyocytes derived from human iPS cells (hiPSC-CM) from the WTCWT line (n = 3, MOI 20).Cells were harvested for RNA extraction at day 4 post-transduction to evaluate MRP1 overexpression via quantitative PCR.MRP1 transcripts were normalized to GAPDH as a measure of overexpression relative to non-transduced (NT) control cells.Statistical significance of differences was calculated using an unpaired t-test (A), or an ordinary one-way ANOVA (B), and the difference compared with NT was calculated with Dunnett's multiple comparison test (** p ≤ 0.01, **** p ≤ 0.0001).(C) Protein was also extracted from transduced hiPSC-CM to determine overexpression of recombinant MRP1 by western blot analysis (n = 3).Data are means ± SEM.

F I G U R E 2
Efflux of 5(6)-carboxy-2 0 ,7 0 -dichlorofluorescein diacetate (CDFA) was increased in hiPSC-CMs overexpressing MRP1.LV.MRP1 was used to transduce hiPSC-CM (MOI 20, n = 3) from the WTC-WT line.Cells were treated with CDFA dye for 2 h at day 4 post-transduction.Dye was then removed and cells were washed two times (at dye removal and again 2 h later).(A) Cells were imaged to visualize GFP fluorescence.(B) Stained cells were harvested, counter-stained for DAPI, then processed for flow cytometry analysis.Histograms show the fluorescence of CDFA in the treated cells.(C) Summary of CDFA accumulated fluorescence intensity from transduced hiPSC-CMs.Data are expressed as fold change over NT 50 μM CDFA for each replicate.Statistical significance of differences was calculated using an ordinary two-way ANOVA (B), and the difference compared with NT was calculated with Sidak's multiple comparison test (** p ≤ 0.01, *** p ≤ 0.001).Data are means ± SEM.
supraphysiological expression of MRP1 in hiPSC-CMs.Transduced cells showed efflux of doxorubicin and robust amelioration of viability loss.Protection was conferred despite residual doxorubicin in MRP1-transduced hiPSC-CMs.Longer exposure to doxorubicin beyond 48 h could be tested to determine whether cardiotoxicity was averted, or simply delayed.The reduction of viability seems to plateau from 50 μM doxorubicin in the LV.LacZ transduced cells.This could indicate the presence of a relatively resistant subpopulation of cells which persists even at high concentrations of doxorubicin.Interestingly, this plateau in viability is also seen in the LV.MRP1 transduced cells.In this case, it is probable that doxorubicin exposure eliminated the NT or low transduction fraction of cells, leaving only the MRP1 positive cells with sufficient transgene expression to be protected.In this study, we propose to protect the heart against anthracycline cardiotoxicity prior to chemotherapy by delivery of a therapeutic gene to the cardiomyocytes.The concept of a transporter gene conferring cardioprotection against anthracyclines has been explored.The multidrug resistance 1 (MDR1) gene has been shown to protect again doxorubicin when overexpressed in mice.14Similar to MDR1, our F I G U R E 3 LV.MRP1 increased the rate of doxorubicin efflux and protected transduced hiPSC-CMs from cell death.LV.MRP1 was used to transduce hiPSC-CM from the SCVI 8 line (MOI 20).(A) Transduced cells in 24-well plates were treated with doxorubicin for 48 h at day 4 post transduction.Doxorubicin fluorescence was quantified using flow cytometry (n = 3) to show the mean fluorescence intensity.(B) Cells were also harvested for RNA extraction at day 4 post-transduction to evaluate MRP1 overexpression via quantitative PCR (n = 3).MRP1 transcripts were normalized to GAPDH as a measure of overexpression relative to non-transduced (NT) control cells.Statistical significance of differences was calculated using an ordinary two-way ANOVA, and the difference between LV.LacZ and LV.MRP1 at each dose of doxorubicin was calculated with Tukey's multiple comparison test (** p ≤ 0.01).(B) Cells were also imaged prior to harvesting to visualize doxorubicin red fluorescence by microscopy in live cells (scale bar = 100 μm).(C) Transduced cells in 96-well plates were treated with doxorubicin for 48 h at day 4 post transduction (n = 8).Viability was then assessed using the MTT assay, with percentage viability expressed as a percentage of the non-injured control (0 μM doxorubicin) of each group.Statistical significance of differences was calculated using an ordinary two-way ANOVA, and the difference compared with LV.LacZ at each dose of doxorubicin was calculated with Sidak's multiple comparison test (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.Data are means ± SEM.