Epigenetic Reexpression of Hemoglobin F Using Reversible LSD1 Inhibitors: Potential Therapies for Sickle Cell Disease.

Sickle cell disease (SCD) is caused by a single nucleotide polymorphism on chromosome 11 in the β-globin gene. The resulting mutant hemoglobin S (HbS) is a poor oxygen transporter and causes a variety of vascular symptoms and organ failures. At birth, the DRED epigenetic complex forms and silences the γ-globin gene, and fetal hemoglobin (HbF, 2 α-, and 2 γ-subunits) is replaced by adult HbA (α2β2) or HbS (α2βs 2) in SCD patients. HbF is a potent inhibitor of HbS polymerization, thus alleviating the symptoms of SCD. The current therapy, hydroxyurea (HU), increases γ-globin and the HbF content in sickle cells but is highly underutilized due to concern for adverse effects and other complications. The DRED complex contains the epigenetic eraser lysine-specific demethylase 1 (LSD1), which appears to serve as a scaffolding protein. Our recently discovered 1,2,4-triazole derivatives and cyclic peptide LSD1 inhibitors promote the upregulation of γ-globin production in vitro without significant toxicity. Herein, we demonstrate that these LSD1 inhibitors can be used to disrupt the DRED complex and increase the cellular HbF content in vitro and in vivo. This approach could lead to an innovative and effective treatment for SCD.

Introduction

Chromatin remodeling is mediated largely through a combination of DNA methylation and histone modifications and is a major regulator of eukaryotic gene expression.[1−6] Histone modifications can include acetylation, phosphorylation, methylation, and a number of other posttranslational modifications (PTMs), resulting in a combination of histone marks collectively known as the histone code.[2] This combination of chromatin marks at a given promoter determines, in part, whether specific genes are in an open/transcriptionally active conformation or a closed/transcriptionally repressed conformation.[2,7,8] In cancer and other diseases, DNA promoter hypermethylation in combination with abnormal histone modifications has been associated with the aberrant silencing of genes.[4,9,10] As a result, multiple chromatin remodeling enzymes have been targeted for the discovery of novel antitumor agents.[1,11,12]

We have concentrated our discovery efforts on the flavin-dependent amine oxidase lysine-specific demethylase 1, (LSD1, also known as KDM1A). The primary function of LSD1 is to remove methyl groups from the activating chromatin mark histone 3 lysine 4 (H3K4). LSD1 is specific for the substrates monomethyl histone 3 lysine 4 (H3K4me) and dimethyl histone 3 lysine 4 (H3K4me2) but is also proposed to demethylate histone 3 lysine 9 (H3K9) when colocalized with the androgen receptor in prostate tumors[13] and has nonhistone protein substrates such as p53 and deoxynucleic acid methyltransferase 1 (DNMT1).[14] A number of effective LSD1 inhibitors have been identified (Figure ) and include tranylcypromine (TCP)-based irreversible inhibitors such as GSK2879552[15] and ORY-1001,[16−18] oligoamines such as verlindamycin[5] and related isosteric ureas and thioureas,[19,20] aralkyl amidoximes,[21] reversible benzohydrazide inhibitors such as SP-2509,[18] and dithiocarbamate-urea hybrid LSD1 inactivators.[22] These inhibitors were developed as antitumor agents and cause varying levels of cytotoxicity in human cells, making them unsuitable for use in diseases where cytotoxicity is not a desired endpoint. Recently, LSD1 has emerged as an important drug target for diseases other than cancer, including neurological disease,[23,24] blood disorders,[6,25] viral infection,[26] diabetes,[27,28] and fibrosis.[29] Thus, there is an unmet medical need for nontoxic LSD1 inhibitors for the treatment of these noncancer disorders.

Figure 1

Structure of the LSD1 inhibitors 1 and 2 (TCP-based), verlindamycin 3, 3,5-diamine-1,2,4-triazoles 4–8, and cyclic peptides 9–11.

At birth, humans express fetal hemoglobin (HbF) composed of 2 α- and 2 γ-globin chains. Within 6 months, the γ-globin gene is epigenetically silenced, and HbF is replaced by HbA, containing 2 α- and 2 β-globin chains. Sickle cell disease (SCD) is caused by an autosomal recessive single-nucleotide polymorphism on chromosome 11 in the β-globin gene that features a Glu6 to Val6 mutation, resulting in the formation of hemoglobin S (HbS).[30] When deoxygenated, HbS becomes insoluble, adheres to endothelial cells, polymerizes, and causes erythrocytes to become fragile and assume a characteristic sickle shape. Erythrocytes containing HbS cause a variety of vascular symptoms (vaso-occlusive episodes, acute chest syndrome, and hemolytic anemia) and organ failures.[31,32] The disease is estimated to occur in 1:300–1:500 African Americans and 1 in 1000 to 1400 Hispanic Americans, and the median age at death is approximately 42 years for men and 48 years for women. The most effective therapy for SCD, the antineoplastic agent hydroxyurea (HU), increases the HbF content in sickle cells by a mechanism that has not been fully elucidated.[33] HbF is a potent inhibitor of the polymerization of deoxyhemoglobin S because neither HbF (α2γ2) nor hybrid tetramers such as α2βSβA and α2βSγF is incorporated into the polymer phase.[34] Thus, enhanced HbF production by HU alleviates the symptoms of SCD. HU is quite efficacious in some individuals but is highly underutilized due to concern for adverse effects, variable levels of HbF induction, and the need for patient monitoring.[35,36] A few agents have been identified that induce γ-globin, but none are superior to HU.[37−39]

The DRED epigenetic complex is a fetal globin gene repressor that silences the γ-globin gene at 6 months of age, and HbF is replaced by adult HbA or HbS in SCD patients.[6,25,40] The DRED complex, first described by Engel et al. in 2007,[41,42] contains LSD1, deoxynucleotide-N-methyltransferase 1 (DNMT1), and the nuclear receptor proteins TR2/TR4,[43] and interacts with corepressors such as NuRD and CoREST. Because DRED controls the expression of γ-globin (and hence HbF), it is considered a target for agents to treat SCD.[40] It has been shown that TCP causes a dose-dependent induction of HbF in erythroid cells, suggesting that the inhibition of LSD1 alleviates SCD.[6] However, TCP causes a multitude of off-target effects, largely because it inhibits the closely related enzymes monoamine oxidase A (MAO-A) and monoamine oxidase B (MAO-B), thus reducing its utility as an SCD treatment. Induction of HbF production in mice has also been observed following treatment with the TCP-based LSD1 inhibitor RN-1 (Figure ),[44−46] but the effect is only moderate, accompanied by the off-target effects of menorrhagia and thrombocytopenia. While there is value in exploiting the TCP scaffold to access novel LSD1 inhibitors for antitumor activity, we have focused our efforts on identifying novel scaffolds to develop potent but reversible inhibitors. There are two main reasons for this strategy: (i) TCP-based inhibitors require metabolic activation to a free-radical species that form an irreversible covalent bond with the FAD+ cofactor. It is likely that this free-radical intermediate could form or react at other sites, which would lead to off-target toxicities. There are acute toxicities associated with TCP-based LSD1 inhibitors, as evidenced by the fact that six clinical trials to evaluate them as treatments for sickle cell disease have been abruptly terminated at an early stage. (ii) It is well established that LSD1 can only demethylate histone methyllysine substrates as a component of an epigenetic complex.[47] The ability of an LSD1 inhibitor to affect chromatin remodeling is thus dependent in part on its ability to disrupt an LSD1-containing epigenetic complex. This activity does not necessarily correlate with IC50 against purified LSD1/CoREST,[48] a finding that underscores the importance of LSD1 as a scaffolding protein.

We have reported multiple chemical scaffolds for inhibition of LSD1 that do not contain a tranylcypromine core structure.[5,21,49−51] Most recently, we have discovered a series of 3,5-diamino-1,2,4-triazoles[51,52] and a series of cyclic peptides[50,53] that are effective inhibitors of LSD1/CoREST in vitro. In light of the role of LSD1 in the DRED epigenetic complex, we evaluated the ability of triazoles 4–8 and cyclic peptides 9–11 (Figure ) to promote the reexpression of the γ-globin gene, thereby increasing the HbF content.

Results and Discussion

3,5-Diamino-1,2,4-triazoles 4, 5, and 6 were selected from previously reported triazole-based LSD1 inhibitors synthesized in our laboratories. This series was identified using a virtual screen strategy[51] followed by development of a synthetic pathway (Scheme S1) and preliminary hit-to-lead optimization.[52] Compounds 7 and 8 were synthesized using an adapted synthetic route to determine whether the chlorine substituent was required for inhibitory activity, as shown in Scheme . Thus, the 1,3-dihalogenated-2-cyano benzene 12 was reduced in the presence of LiAlH4 to yield the corresponding primary amine 13. Compound 13 was then appended to dimethyl cyanodithioiminocarbonate 14 to afford intermediate 15, which was cyclized in the presence of hydrazine to form 16. Intermediate 16 was then coupled to either naphthalen-1-ol or 4-phenylphenol to afford compounds 7 or 8, respectively. Unexpectedly, conversion of 16 to the final products 7 and 8 yielded the corresponding dechlorinated product. This effect may be due to stabilization of the aromatic ring via anchimeric assistance that promotes the elimination of the chlorine prior to attack of the phenol on fluorine. The complete synthetic details for compounds 4–8 and the associated analytical data are described in the Supporting Information. Compounds 7 and 8 were also found to inhibit the polyamine catabolic enzyme spermine oxidase (SMOX).[52]

Scheme 1

Synthetic Route to Dechlorinated Compounds 7 and 8

Cyclic peptides 9–11 were chosen based on their LSD1 inhibitory activity from a rationally designed library[50] that was optimized via alanine scanning.[53] The syntheses of peptides 9–11 have been previously reported,[50,53] and the associated analytical data are described in the Supporting Information. The ability of compounds 2–11 to inhibit LSD1/CoREST in vitro is outlined in Table .

Table 1

Inhibition of LSD1 by Compounds 4–11

compound	LSD1 IC50 (μM)
2	0.016
3	13.0[57]
4	1.19[51]
5	0.230[52]
6	0.190[52]
7	ND (80.2% inhibition at 10 μM)[52]
8	ND (50.0% inhibition at 10 μM)[52]
9	ND (71% inhibition at 500 nM)[53]
10	ND (86% inhibition at 500 nM)[53]
11	0.136[53]

The individual components of the DRED complex are commonly found in epigenetic complexes at multiple gene promoter sites. As such, it is not feasible to measure whether individual components released from an epigenetic complex arise specifically from DRED. However, the observed reexpression of γ-globin could not occur unless DRED was disrupted. In a related study, we performed a coimmunoprecipitation study in isolated rat cardiomyocytes to determine whether LSD1 inhibitors such as 3 and 4 were capable of disrupting epigenetic complexes (Figure ). Importantly, compound 4, but not compounds 3 or 17, effectively disrupt an LSD1-containing epigenetic complex, as determined by the absence of CoREST in the pull-down mixture. These data indicate that not all inhibitors of LSD1 are capable of disrupting such a complex and underscore the importance of LSD1 as a scaffold protein. CD34+ erythrocyte progenitor cells require transformation to a nonhomogeneous mixture of immature erythroblasts prior to measurement of the HbF content; as such, we are now developing an assay to perform similar coimmunoprecipitation experiments in CD34+-derived cells.

Figure 2

Coimmunoprecipitation analysis of the disruption of an LSD1:CoREST complex in primary rat cardiomyocytes following treatment with 1.0 μM 3, 1.0 μM 4, and 2.0 μM 17. Compounds 3 and 17 are LSD1 inhibitors previously described by our group.[49,51] Lysates were pulled down with the LSD1 antibody and immunoblotted for CoREST. In the absence of the inhibitor, the LSD1 antibody also pulled down HDAC2 from the epigenetic complex. These data are representative of three separate experiments.

The K562 human myelogenous leukemia cell line is widely used as a model for studying in vitro erythropoiesis and hemoglobin gene regulation and, as such, is well suited to study induction of γ-globin in vitro due to its ability to produce fetal hemoglobin and F-cells.[54−56] We first determined the toxicity of representative LSD1 inhibitors tranylcypromine, verlindamycin 3, 3,5-diamino-1,2,4-triazole 4, and cyclic peptide 11 in K562 cells using a standard MTS cell viability assay (Figure ). Tranylcypromine and 4 exhibited minimal cytotoxicity in the K562 cell line, while cyclic peptide 11 was moderately cytotoxic, and verlindamycin 3 was cytotoxic with an IC50 value near 10 μM, as previously reported.[57]

Figure 3

Cytotoxicity of tranylcypromine, 3, 4, and 11 in cultured K562 human myelogenous leukemia cells.

To determine whether LSD1 inhibitors 4–11 could promote the expression of γ-globin, we measured the γ -globin content in treated K562 cells using RT-qPCR. There is a great deal of disparity among the LSD1 IC50 values for the agents used in this study, as shown in Table . To make a fair comparison, we dosed these agents at concentrations at or near their IC50 values. We felt that choosing concentrations near the IC50 would even out the LSD1 inhibitory effects across the spectrum of IC50 values and thus result in a more meaningful comparison. For example, dosing the antitumor compound 2 (LSD1 IC50 = 0.016 μM) at 1.0 μM would be far above the amount required to produce 100% inhibition of the enzyme and would kill all of the cells. This would not provide a meaningful comparison to compound 4, which has an IC50 value of 1.19 μM with no cytotoxicity. The results of these experiments are shown in Figure .

Figure 4

Increases in γ-globin mRNA expression in the K562 human myelogenous leukemia cell line, as measured by RT-qPCR. Each data point is the average of three separate determinations ± SEM. Compared to control, NS, not significant; P < 0.0015; ***P < 0.0005; ****P < 0.0001.

Hemin, a known inducer of hemoglobin production in erythroid cells, produced a 2.4-fold increase in γ-globin production, while HU, the current standard of therapy, produced a 2.3-fold increase. Importantly, the potent and specific tranylcypromine-based LSD1 inhibitor GSK-LSD1 2 promoted a 1.3- and 1.4-fold increase in γ-globin expression at 16 and 32 nM, respectively. This observation suggests, as we have hypothesized, that the effect of LSD1 inhibitors on gene expression can vary, depending on the specific epigenetic complex in which LSD1 is a component. The polyamine-based LSD1 inhibitor 3 produced a 4.1-fold induction. In the triazole series, 4 produced a 1.75-fold induction of γ-globin at 1.5 μM. By contrast, the homologous analogue 6 promoted a 12.6- and 10.6-fold expression of γ-globin at concentrations of 0.5 and 1.0 μM, respectively. In addition, the cyclic peptides 9, 10, and 11 were effective promoters of γ-globin expression, eliciting 3.4-, 5.8-, and 7.3-fold increases, respectively.

In previous studies, we have observed that responses in different cell lines to epigenetic modulators are cell type-specific. As such, we chose to measure the induction of γ-globin in a more relevant cell line. We monitored γ-globin expression production in human CD34+ erythroid progenitor cells using RT-qPCR. Compounds 3 and 9–11 were not advanced to this stage due to potential toxicities and the difficulties associated with peptide-based small molecules in cell-based systems. In addition, two more potent inhibitors in the series, 7 and 8, were also evaluated. The results of these studies are shown in Figure . Hemin, HU, and GSK-LSD1 all promoted increases in γ-globin that were analogous to levels observed in the K562 line. However, compounds 4–8 were all effective at enhancing γ-globin expression by 2.4- to 5.4-fold at concentrations between 1 and 50 μM. As before, compound 6 was the most effective agent, promoting a 5.4-fold increase in γ-globin at 1 μM and 4.3-fold at 10 μM. The reduction in γ-globin production at 10 μM could be due to mild cytotoxicity of 6 at this higher dose level.

Figure 5

Increases in γ-globin mRNA expression in human CD34+ erythroid progenitor cells, as measured by RT-qPCR. Each data point is the average of three separate determinations + SEM. Compared to control, NS, not significant; *P < 0.03; **P < 0.0015; ***P < 0.0005; ****P < 0.0001.

Compounds 4–8 were also evaluated for the ability to increase β-globin levels, as measured by RT-qPCR (Figure ). Hemin promoted a 1.4-fold elevation of β-globin, while HU had no effect. GSK-LSD1 2 also did not increase β-globin levels and in fact promoted a marked decrease at both 20 and 50 nM. Compound 4 caused a 2.2-fold increase in β-globin, while 5–8 promoted modest increases between 1.5- and 1.8-fold.

Figure 6

Increases in β-globin mRNA expression in human CD34+ erythroid progenitor cells, as measured by RT-qPCR. Each data point is the average of three separate determinations + SEM. *P < 0.03; **P < 0.0015; ***P < 0.0005.

The ability of a small molecule to promote the expression of γ-globin is best modeled in vivo using the Townes murine model for SCD. The Townes transgenic mouse is a knock-in model in which the entire human hemoglobin gene construct has been knocked into a mouse with a mixed B6/129 genetic background. This transgenic mouse exclusively expresses human α-, β-, and γ-globin, effectively mimics human postnatal hemoglobin class switching from γ-globin to mutant or normal β-globin, and pups express up to 50% HbF (γ-globin) at birth and switch to human adult HbA (β-globin) by 4 weeks of extrauterine life.[58,59] As such, the Townes mouse provides a humanized sickle cell mouse model and controls (expressing normal adult β-globin) and thus is a suitable model for the proposed experiments.

In preliminary in vivo studies in Townes SS mice, we were able to detect the presence of F-cells and F-reticulocytes by flow cytometry using the Cytoflex platform and 5–10 μL of blood taken from the tail vein. Compounds 4 and 8 at 10 mg/kg promoted significant increases in HbF-containing erythrocytes following 2 weeks of treatment (Figure A). There was a corresponding increase in HbF-reticulocytes (30–37% for 6, 30–33% for 8, not shown). HPLC analysis of these blood samples revealed an 87% increase (relative to baseline/pretreatment levels) in the HbF content after 2 weeks at 10 mg/kg (Figure B).

Figure 7

(A) Two-week increase in the percentage of F-cells for compounds 4 and 8 at 10 mg/kg. Each data point is the average of three animals + SEM. (B) HPLC trace showing that, relative to baseline or pretreatment (black line), compound 8 promoted an 87% increase in γ-globin in the Townes mouse model following 2 weeks of treatment.

Conclusions

In conclusion, we have identified a series of 3,5-diamino-1,2,4-triazoles and a series of cyclic peptides that are potent inhibitors of the histone demethylase LSD1. Partial hit-to-lead optimization has been performed on both of these series of compounds. Because the LSD1 inhibitors described above have low toxicity in normal mammalian cells, we evaluated these agents for their ability to disrupt the LSD1-containing DRED epigenetic complex, which is responsible for the silencing of the γ-globin gene and hence HbF production at 6 months of age. Our results demonstrate that these compounds are effective at promoting the reexpression of γ-globin and hence HbF, presumably by disrupting the DRED complex and allowing reexpression of the γ-globin gene. We have demonstrated these effects in the K562 cell line, in normal human CD34+ erythroid progenitor cells, and in vivo. Although more extensive toxicology studies are required, our preliminary results suggest that our inhibitors will be better tolerated than previously reported tranylcypromine-based LSD1 inhibitors that have not been successful in the clinic for the treatment of SCD. The synthesis and evaluation of additional analogues in both the triazole and cyclic peptide series are ongoing concerns in our laboratories.

Experimental Section

Chemistry

All reagents and dry solvents were purchased from Aldrich Chemical Co. (Milwaukee, WI), Sigma Chemical Co. (St. Louis, MO), VWR (Radnor, PA), or Thermo Fisher Scientific (Chicago, IL) and were used without further purification, except as noted below. Triethylamine was distilled from potassium hydroxide and stored in a nitrogen atmosphere. Dry methanol, ethyl acetate, tetrahydrofuran, dimethyl formamide, and hexane were prepared using a glass contour solvent purification system (Pure Process Technology, LLC, Nashua, NH). Preparative scale chromatographic procedures were carried out using a CombiFlash Rf200 chromatography system (Teledyne ISCO, Lincoln, NE) fitted with silica gel 60 cartridges (230–440 mesh). Thin layer chromatography was conducted on Merck precoated silica gel 60 F-254. Ion exchange chromatography was conducted on Dowex 1 × 8 200 anion exchange resin.

All 1H and 13C NMR spectra were recorded on a Bruker Avance 600 MHz spectrometer, and all chemical shifts are reported as δ values referenced to TMS or DSS. Splitting patterns are indicated as follows: s, singlet; d, doublet; t, triplet; m, multiplet; br, broad peak. In all cases, 1H NMR, 13C NMR, and MS spectra were consistent with assigned structures. Mass spectra were recorded by LC/MS on a Waters AutoPurification liquid chromatography with a model 3100 mass spectrometer detector. Prior to biological testing procedures, all compounds were determined to be >95% pure by UPLC chromatography (95% H2O/5% acetonitrile to 20% H2O/80% acetonitrile over 10 min) using a Waters Acquity H-series ultrahigh-performance liquid chromatograph fitted with a C18 reversed-phase column (Acquity UPLC BEH C18 1.7 M, 2.1 × 50 mm). Compounds 4–11 were synthesized according to the general procedures described in the Supporting Information.

Enzyme Assay

Compounds were evaluated for the ability to inhibit recombinant LSD1/CoREST using a commercially available assay kit (#700120, Cayman Chemical, Ann Arbor, MI). The substrate and all compounds were incubated in assay buffer from 30 min up to 4 h at 37 °C, as described in the commercial protocol. The volume of each reaction well was 50 μL, containing 5 μL of a 200 μM solution of substrate peptide and 20 μL of a 15 ng/μL enzyme solution. All compounds were diluted in 1% DMSO with assay buffer to a final volume of 50 μM. Fluorescence was measured at the recommended wavelengths of kex = 530 nm, kem = 590 nm. IC50 determinations were performed using serial dilutions at 250, 125, 62.5, 31.25, 6.25, 3.125, 1.563, 0.781, and 0.390 μM. All blanks contained 1% DMSO to determine any solvent effects.

Cell Culture

K562 (human myelogenous leukemia) cells were purchased from ATCC. Cells were cultured in Iscove’s modified Dulbecco’s medium containing 10% (v/v) fetal bovine serum and 5% penicillin and streptomycin, as previously described.[53] All cultures were grown at 37 °C in a humidified environment containing 5% CO2.

CD34+ cells were expanded in H3000 media supplemented with CC100 (Stem Cell Technologies) for 4 days at a density of 105 cells/mL prior to use for compound evaluation. Cells were treated for 24, 48, or 72 h with an appropriate concentration of each analogue.

Cell Viability Assay

All cell lines were purchased from ATCC (Manassas, VA). For the (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxy-methoxy-phenyl)-2-(4-sulfo-phenyl)-2H-tetrazolium) (MTS) reduction assay, 2000 cells per well were seeded in 50 μL of a complete medium per well of a 96-well plate and the cells were allowed to attach overnight at 37 °C in a 5% CO2 atmosphere. The medium was aspirated, and cells were treated with 100 mL of a fresh medium containing appropriate concentrations of each inhibitor to be tested. The cells were incubated for 72 h at 37 °C in 5% CO2 after which 20 mL of the MTS reagent solution (Promega CellTiter 96 Aqueous One Solution Cell Proliferation Assay) was added to the medium. The cells were incubated for another 2 h at 37 °C, and absorbance was measured at 490 nm on a SpectraMax M5 instrument (Molecular Devices) equipped with SOFTmax PRO 4.0 software to determine cell viability. A reference wavelength of 690 nm was used to subtract the background. The percentage of cell death was calculated by the following equation: % cell death = (abs control – abs sample)/abs control × 100. A dose-response curve was constructed for each inhibitor, and each data point was the average of three determinations obtained during a single experiment ± SEM. IC50 values were calculated using the GraphPad Prism 5 software package (GraphPad, San Diego, California).

RT-qPCR

Media were rinsed with PBS and cells were lysed using the TriZol reagent (Invitrogen; cat #15596026). mRNA was isolated according to the manufacturer’s protocols and purity-confirmed using a Nanodrop-1000 spectrophotometer (Thermo Fisher Scientific). Quantitative reverse-transcription real-time polymerase chain reaction (qRT-PCR) was run using a high-capacity cDNA reverse transcription kit (Applied Biosystems; cat# 4368814) followed by a TaqMan fast advanced master mix (Applied Biosystems; cat #4444557) with TaqMan gene expression assay primers (Applied Biosystems, listed below) using a StepOne Plus instrument (Thermo Fisher Scientific). TNF-α, IL-6, and the internal control GAPDH were then quantitated for each sample in triplicate. Results are reported as fold change (2–ΔΔCT).

Primary Rat Cardiomyocyte Isolation

Primary male Sprague–Dawley rat cardiomyocytes were isolated via a hanging heart preparation using enzymatic digestion, as previously described.[51] In brief, rats were euthanized with 5% isoflurane vaporized in 100% O2. The heart was retrogradely perfused with collagenase. The cardiomyocytes were plated on six-well culture trays that were coated with laminin at an initial plating density of 1.5 × 105 cells/well. After overnight incubation, the cardiomyocytes were rinsed and maintained in a serum-free medium.

Coimmunoprecipitation

Primary rat cardiomyocytes were isolated as described earlier and treated with varying concentrations of test compounds. Cells were lysed and scraped with IP lysis buffer (20 mM Tris–Cl, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM β-glycerol, and 2.5 mM Na pyrophosphate). Dynabeads (Life Biosciences) were added (1:10 beads:lysate, v/v) and incubated at 4 °C for 1 h to preclear. The mixture was centrifuged (5000 rpm), and the supernatant was saved. An antibody (2–5 μg) for the protein of interest was added and rocked overnight in a cold room. Dynabeads (20 μL) were added and incubated at room temperature for 1 h. The mixture was centrifuged at 5000 RPM for 1 min.The supernatant was decanted, and the pellet was washed three times with IP lysis buffer with 0.1% Triton X-100. Samples were then analyzed by SDS-PAGE and immunoblotted.