Research ArticleCancer

Tissue-specific regulation of p53 by PKM2 is redox dependent and provides a therapeutic target for anthracycline-induced cardiotoxicity

See allHide authors and affiliations

Science Translational Medicine  06 Feb 2019:
Vol. 11, Issue 478, eaau8866
DOI: 10.1126/scitranslmed.aau8866

Killing two birds with one protein

Anthracyclines, a class of common chemotherapy drugs, frequently cause cardiac damage. This cardiotoxicity is associated with activation of p53, a protein that stimulates apoptosis in tumors and in other organs such as the heart, causing nonspecific tissue injury. Saleme et al. found that pyruvate kinase muscle 2 (PKM2) directly interacts with p53. The tetrameric form of PKM2 is preferentially oxidized in the heart, where it suppresses apoptosis. Conversely, the same tetramer enhances the activity of p53 in low oxidation environments such as tumors. The authors identified a compound that stabilizes tetrameric PKM2, protecting cardiomyocytes while facilitating cancer cell killing by anthracyclines.

Abstract

Chemotherapy-induced cardiotoxicity (CIC) is a common clinical problem that compromises effective anticancer therapies. Many chemotherapeutics (including anthracyclines, such as doxorubicin) induce the proapoptotic transcription factor p53 in the tumor and nonspecifically in the heart, promoting heart failure. Although inhibition of p53 shows benefit in preclinical heart failure models, it would not be an attractive adjuvant therapy for CIC, because it would prevent tumor regression. A p53-targeting therapy that would decrease chemotherapy-induced apoptosis in the myocardium and, at the same time, enhance apoptosis in the tumor would be ideal. Here, we propose that differences in oxygen tension between the myocardium and the tumor could provide a platform for redox-dependent tissue-specific therapies. We show by coimmunoprecipitation and mass spectrometry that the redox-regulated pyruvate kinase muscle 2 (PKM2) directly binds with p53 and that the redox status of cysteine-423 of tetrameric (but not monomeric) PKM2 is critical for the differential regulation of p53 transcriptional activity. Tetrameric PKM2 suppresses p53 transcriptional activity and apoptosis in a high oxidation state but enhances them in a low oxidation one. We show that the oxidation state (along with cysteine-423 oxidation) is higher in the heart compared to the tumor of the same animal. Treatment with TEPP-46 (a compound that stabilizes tetrameric PKM2) suppressed doxorubicin-induced cardiomyocyte apoptosis, preventing cardiac dysfunction, but enhanced cancer cell apoptosis and tumor regression in the same animals in lung cancer models. Thus, our work suggests that redox-dependent differences in common proteins expressed in the myocardium and tumor can be exploited therapeutically for tissue selectivity in CIC.

INTRODUCTION

The redox state of a cell is important for the regulation of many fundamental cellular processes, including cell survival or apoptosis (1). An important and emerging redox-regulated posttranslational modification of proteins is the oxidation of cysteine residues, which are implicated in many cell signaling events (2). Although tissue-specific redox differences, along with the response to therapy, have been previously described (3), it remains unknown whether redox differences between tissues (and the subsequent oxidation/reduction of proteins) could provide a therapeutic platform for tissue-specific therapies. The need for such a selective approach is especially evident in cancer, where many effective therapies have adverse effects on normal tissues (4).

A classic example of nonselective cancer therapeutics is chemotherapy-induced cardiotoxicity (CIC), a condition that results from the use of many chemotherapeutic agents, including anthracyclines and tyrosine kinase inhibitors (5, 6). Because the survival of patients with cancer has increased over the past two decades, a larger portion of cancer survivors are now presenting with cardiovascular diseases later in life (7), suggesting that chemotherapeutics can promote myocardial dysfunction even years after chemotherapy is completed. Although the appreciation of CIC has resulted in the establishment of clinical cardio-oncology programs (5), there are still no therapies to prevent or limit CIC. The main problem with current chemotherapeutics is the nonspecific induction of apoptosis in noncancer tissues, including the myocardium (8). Although most normal tissues, which are composed of replicating epithelial cells, can recover from chemotherapy (9), the cardiomyocytes are terminally differentiated and have limited replicative capacity (10), making the heart more susceptible to long-term dysfunction. This problem highlights the difficulties in designing appropriate CIC therapies that would protect the myocardium from chemotherapy-mediated apoptosis (and subsequently heart failure) without hindering tumor apoptosis and tumor regression. Although an ideal strategy would be to target a protein that is exclusively expressed in the tumor, an alternative approach may be to target a protein expressed in both the myocardium and the tumor but is regulated differently in a tissue-specific manner. For example, the myocardium is exposed to higher levels of partial pressures of oxygen (which promotes protein cysteine oxidation) compared to the tumor (11). Therefore, targeting redox-sensitive proteins that may inhibit chemotherapy-induced apoptosis, when oxidized in the myocardium, could provide therapeutic selectivity for the myocardium without compromising tumor regression.

Many effective chemotherapeutics, including anthracyclines, induce and activate the proapoptotic transcription factor p53 in both cancer cells, resulting in tumor regression (12), and cardiomyocytes, resulting in heart failure (13). Inhibition of p53 is beneficial in a variety of heart failure models (1315), but this would not be a valid approach for preventing CIC because it would also promote tumor growth. Identifying specific redox-sensitive proteins that could interact with and differentially regulate p53 transcriptional activity and apoptosis in the myocardium (a high oxidation environment) compared to the tumor (a low oxidation environment) could provide therapeutic approaches to prevent CIC, yet simultaneously enhance chemotherapy-mediated tumor regression.

Here, we show that the redox-regulated pyruvate kinase muscle 2 (PKM2) can bind with p53 and differentially regulate its transcriptional activity, depending on the redox status of PKM2. Oxidized [cysteine-423 (C423)] tetrameric (but not monomeric) PKM2 can suppress p53 transcriptional activity and apoptosis, whereas reduced tetrameric PKM2 can enhance p53 transcriptional activity and apoptosis. We found that the myocardium is much more oxidized compared to the tumor and that PKM2 is preferentially oxidized in the myocardium (on C423) compared to the tumor. Treatment with TEPP-46 (a compound that stabilizes tetrameric PKM2) prevented doxorubicin-induced p53 transcriptional activity and apoptosis in cardiomyocytes and in vivo in the myocardium, preventing cardiac dysfunction, but at the same time and in the same animal, it enhanced doxorubicin-induced p53 transcriptional activity and apoptosis in lung cancer cells, improving lung tumor regression. Thus, this tissue-specific differential regulation of p53 transcriptional activity and apoptosis by tetrameric PKM2 could provide a therapeutic strategy against chemotherapy-induced cardiac dysfunction. Our work shows that tissue-specific redox regulation of drug targets can provide a platform for therapeutic selectivity.

RESULTS

PKM2 (but not PKM1) can bind to p53 and inhibit its transcriptional activity

To identify candidate redox-sensitive proteins that could modulate p53 transcriptional activity, we performed coimmunoprecipitation of p53 in CRL-2321 cells [a cell line that expresses nuclear wild-type (WT) p53 at baseline; Fig. 1A], followed by mass spectrometry. We identified PKM, a redox-regulated metabolic enzyme (16) that is found in both the cytoplasm and the nucleus (17), as a candidate binding partner for p53 (Fig. 1A). PKM has two main isoforms, and we identified PKM2, but not PKM1, as the candidate PKM isoform interacting with p53 in A549 and HTB177 lung cancer cells (Fig. 1B and fig. S1A). Lysine-433 (K433) of PKM2 has previously been shown to be important for its direct interactions with β-catenin and is specific to PKM2 [PKM1 encodes for glutamic acid (E) at position 433] (17); therefore, we investigated whether K433 was important for PKM2 binding to p53. Mutation of K433 to E resulted in almost complete loss of PKM2 binding to p53 (Fig. 1C and fig. S1B).

Fig. 1 PKM2 interacts with p53, and monomeric PKM2 can inhibit its transcriptional activity.

(A) CRL-2321 mammary epithelial cells have high expression of endogenous nuclear p53 (red), imaged using confocal microscopy. Immunoprecipitation of p53 followed by mass spectrometry identified five unique peptides for PKM. Scale bar, 20 μm. (B) Immunoprecipitation (IP) of Flag in Flag-PKM1/PKM2 expressing A549 non–small cell lung cancer epithelial cells shown using immunoblots. IgG, immunoglobulin G. (C) Immunoprecipitation of p53 from A549 cells expressing WT or K433E PKM2 shown using immunoblots. (D) To generate PKM-deficient cells, we inserted a tGFP gene, which also contains a phosphoglycerate kinase (PGK)–driven puromycin (PM) resistance gene, downstream of the PKM promoter using CRISPR-Cas9 homologous repair, such that induction of the endogenous PKM promoter would result in the expression of tGFP instead of PKM. The red rectangle after each gene denotes a transcriptional stop signal. (E) Immunoblots showing the amounts of PKM1, PKM2, and tGFP in parental and PKM-deficient cells. Actin was used as a loading control. (F) p21 and PUMA mRNA expression was measured using quantitative reverse transcription polymerase chain reaction (qRT-PCR) in parental, PKM-deficient, and PKM-deficient A549 cells expressing PKM2. 18S was used as a housekeeping gene (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to parental, #P < 0.05 compared to PKM deficient). (G) p21 and PUMA mRNA expression is measured using qRT-PCR in parental, PKM-deficient, and PKM-deficient A549 cells expressing R399E PKM2. 18S was used as a housekeeping gene (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to parental, #P < 0.05 compared to PKM deficient). (H) Firefly luciferase for p53 activity in A549 parental and PKM-deficient cells expressing PKM1 or various PKM2 mutants. Renilla luciferase was used as a transfection control (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to parental, #P < 0.05 compared to PKM deficient).

To investigate the functional importance of PKM2 and p53 interactions, we generated a PKM-deficient cell line. We used CRISPR-Cas9 genome editing to insert a turbo green fluorescent protein (tGFP) gene with a transcriptional stop codon downstream of the PKM promoter (Fig. 1D) and generated PKM1- and PKM2-deficient cell lines that express tGFP instead (Fig. 1E). In PKM-deficient cells, we observed an increase in the expression of the p53 target genes CDKN1A (p21) and PUMA (18) compared to parental cells (Fig. 1F). The increase in p21 and PUMA expression was also confirmed in an acute model of PKM2 knockdown [via small interfering RNA (siRNA)] in A549 cells compared to control-treated cells (fig. S1C). Furthermore, overexpression of Flag-PKM2, but not Flag-PKM1 or Flag-K433E-PKM2 in PKM-deficient cells, reversed the increase in p21 and PUMA (Fig. 1F and fig. S1D), suggesting that PKM2 can inhibit p53 transcriptional activity. PKM2 can be present as a monomer, dimer, or tetramer (19), and a recent study showed that monomeric and tetrameric PKM2 can bind with p53 (20). To address the role of monomeric PKM2 in regulation of p53 transcriptional activity, we used the R399E mutant form of PKM2, which is unable to effectively form a tetramer (and remains primarily as a monomer) (21). We found that similar to WT PKM2, R399E-PKM2 was also able to reverse the increase in p53 transcriptional activity in PKM-deficient cells to levels similar to parental cells (Fig. 1G). To provide further evidence that PKM2 is important in the regulation of p53 transcriptional activity, we used a p53-specific luciferase assay. Similar to endogenous p53 transcriptional targets, we found that PKM-deficient cells had increased p53 luciferase activity compared to parental cells and that overexpression of Flag-PKM2 or Flag-R399E-PKM2, but not Flag-PKM1 or Flag-K433E-PKM2 in PKM-deficient cells, reversed the increase in p53 luciferase activity (Fig. 1H). These data suggest that monomeric PKM2 can bind and repress p53 transcriptional activity.

Tetrameric PKM2 inhibits p53 transcriptional activity and apoptosis in an oxidized environment but enhances p53 transcriptional activity and apoptosis in a reduced environment

Because K433 is localized on the periphery of PKM2, away from the monomer-monomer interacting site (fig. S2A), it is plausible that tetrameric PKM2 can bind with p53 (20). To investigate the effects of tetrameric PKM2 on p53 binding and transcriptional activity, we first treated A549 cells with TEPP-46, a small molecule that stabilizes the tetrameric form of PKM2 [by promoting monomer-monomer interactions (19); fig. S2, B and C], and found that TEPP-46 enhanced the binding between PKM2 and p53 (fig. S2D). We next treated A549 cells with TEPP-46 in the presence of the p53 stabilizer nutlin-3a. As expected, nutlin-3a increased p53 luciferase activity compared to vehicle-treated cells (fig. S2E); however, TEPP-46 further increased nutlin-3a–mediated p53 transcriptional activity (fig. S2E). We observed similar increases with TEPP-46 and nutlin-3a for endogenous p53 transcriptional targets [p21-, PUMA-, and TP53-induced glycolysis and apoptosis regulator (TIGAR) expression] as well (fig. S2F). In addition, treatment with DASA-58, another compound that increases the tetrameric form of PKM2 (19), also increased nutlin-3a–induced p53 transcriptional activity, similar to TEPP-46 in A549 cells (fig. S2G). Together, these data suggest that tetrameric PKM2 can bind p53 and enhance its transcriptional activity.

Because PKM2 is redox regulated (16), we studied whether oxidation of tetrameric PKM2 is important in modulating p53 transcriptional activity. We used the NADH (reduced form of nicotinamide adenine dinucleotide) quinone dehydrogenase 1 (NQO1)–mediated two-electron reduction system to generate endogenous H2O2 (22, 23), a substrate for cysteine oxidation (Fig. 2A) (2). The advantage of this system is that H2O2 can be synthesized in the same compartments as PKM2, the cytoplasm and the nucleus, because NQO1 is also present in these same compartments (Fig. 2B; see fig. S2H for secondary-only staining). Treatment of A549 cells [which have high expression of NQO1 (24)] with the NQO1 substrate β-lapachone increased both cytoplasmic and nuclear reactive oxygen species (ROS) production, along with a parallel and predicted increase in cysteine oxidation, measured by a roGFP-Orp1 cysteine oxidation assay (Fig. 2, C and D). In A549 cells overexpressing Flag-p53, TEPP-46 further increased p21 expression, but in the presence of β-lapachone, TEPP-46 prevented the increase in p21 expression (Fig. 2E).

Fig. 2 Oxidized or reduced tetrameric PKM2 can suppress or enhance p53 activity, respectively.

(A) Schematic representation on the left shows the mechanism for NQO1-mediated reduction of a quinone (structure A) to a hydroquinone (structure B). The spontaneous reaction of a hydroquinone with molecular oxygen results in the formation of the more “stable” quinone, along with the synthesis of H2O2. Right: Compounds that are substrates for NQO1 (containing quinone groups identified by the black arrowheads) include β-lapachone, mitoxantrone, ametantrone, and pixantrone. The anthracene bisantrene does not contain a quinone group and is therefore not a substrate for NQO1 but is structurally similar to anthraquinones. (B) PKM2 (purple) and NQO1 (red) localization in A549 cells using confocal microscopy. The merged panel includes the nuclear stain 4′,6-diamidino-2-phenylindole (DAPI; blue). Scale bar, 20 μm. (C) ROS concentrations are measured in A549 cells treated with the NQO1 substrate β-lapachone using CellROX (green) and confocal microscopy. The nuclear stain Hoechst is blue. Scale bar, 20 μm. (D) Cysteine oxidation is measured in A549 cells treated with β-lapachone as assessed by an increase in the excitation of oxidized GFP at 408 nm compared to 488-nm excitation spectra. The oxidation of GFP by β-lapachone was presented as a ratio of 408 (oxidized GFP)/488 (reduced GFP). Confocal images show a representative cell at t = 0 and 100 min, where the 408/488 signal intensity (in red) is calculated by the confocal Zen software. The white arrowheads indicate two different nuclei, and the yellow arrowheads indicate the cytoplasm (n = 18 different cells; *P < 0.05 compared to pretreatment). Scale bar, 20 μm. (E) p21 mRNA expression was assessed by qRT-PCR in A549 cells overexpressing Flag-p53 and treated with TEPP-46 and β-lapachone. 18S was used as a housekeeping gene (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to empty vector, **P < 0.05 compared to Flag-p53, #P < 0.05 compared to Flag-p53 and β-lapachone group). Immunoblot shows expression of Flag-p53 on the right. Actin was used as a loading control. (F) Flag immunoprecipitation of Flag-PKM2–expressing A549 cells treated with vehicle, mitoxantrone, or bisantrene showing the extent of PKM2 cysteine sulfonation (Cys-SO3) using immunoblots. (G) PUMA mRNA expression, as well as p53 and cleaved (*) caspases 7 and 3 (Cas7 and Cas3), was measured using qRT-PCR or immunoblots, respectively, in A549 cells treated with mitoxantrone and/or TEPP-46 (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle, #P < 0.05 compared to mitoxantrone). 18S or actin was used as controls for qRT-PCR or immunoblots, respectively. (H) PUMA mRNA expression, as well as p53 and cleaved caspases 7 and 3, was measured using qRT-PCR or immunoblots, respectively, in A549 cells treated with ametantrone and/or TEPP-46 (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle, #P < 0.05 compared to ametantrone). 18S or actin was used as controls for qRT-PCR or immunoblots, respectively. (I) PUMA mRNA expression, as well as p53 and cleaved caspases 7 and 3, was measured using qRT-PCR or immunoblots, respectively, in A549 cells treated with pixantrone and/or TEPP-46 (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle, #P < 0.05 compared to pixantrone). 18S or actin was used as controls for qRT-PCR or immunoblots, respectively. (J) PUMA mRNA expression, as well as p53 and cleaved caspases 7 and 3, was measured using qRT-PCR or immunoblots, respectively, in A549 cells treated with bisantrene and/or TEPP-46 (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle, #P < 0.05 compared to bisantrene). 18S or actin was used as loading controls for qRT-PCR or immunoblots, respectively. (K) Firefly luciferase for p53 activity in A549 cells treated with mitoxantrone or bisantrene and/or TEPP-46 (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle, #P < 0.05 compared to mitoxantrone, ##P < 0.05 compared to bisantrene).

To address whether oxidation of tetrameric PKM2 can inhibit p53-mediated apoptosis, we used a number of anthraquinone compounds that not only are substrates of NQO1 [and therefore can generate endogenous H2O2 (22)] but also intercalate with DNA to induce p53 and apoptosis [via the induction of PUMA (18)]. We compared them to bisantrene, a similar compound that induces p53-mediated apoptosis but is not a substrate for NQO1 and therefore does not generate large amounts of ROS (Fig. 2A). Treatment of cells with the anthraquinones mitoxantrone, ametantrone, and pixantrone increased PKM2 oxidation [as assessed by PKM2 cysteine sulfonation, an irreversible oxidation modification of cysteine (2)], compared to vehicle or bisantrene (Fig. 2F and fig. S2I), and did not change PKM2 oligomerization (fig. S2J). Whereas TEPP-46 prevented mitoxantrone-, ametantrone-, and pixantrone-mediated PUMA expression and apoptosis (assessed by cleaved caspases 7 and 3; Fig. 2, G to I, red box), it enhanced bisantrene-mediated PUMA expression and apoptosis (Fig. 2J, red box). Similar results were also observed on a p53-specific luciferase assay, with TEPP-46 preventing mitoxantrone-mediated p53 transcriptional activity but increasing bisantrene-mediated p53 transcriptional activity (Fig. 2K). These data suggest that tetrameric PKM2 can differentially regulate p53 transcriptional activity depending on the oxidation state of a cell.

C423 of tetrameric PKM2 is critical for the differential regulation of p53 transcriptional activity and apoptosis

To confirm that oxidation of tetrameric PKM2 is involved in inhibition of p53 transcriptional activity, we mutated all 10 cysteine residues of PKM2 to serine (S; an amino acid that cannot be further oxidized; fig. S3A) before overexpression in a PKM-deficient cell line and treatment with mitoxantrone and/or TEPP-46. We identified C423 as the candidate residue important for oxidized tetrameric PKM2-mediated inhibition of p53 transcriptional activity, because TEPP-46 decreased mitoxantrone-mediated PUMA expression in all mutants except C423S (fig. S3B, blue box). In addition, TEPP-46 was able to decrease mitoxantrone-mediated apoptosis in WT but not C423S-PKM2–expressing cells (Fig. 3A, red box).

Fig. 3 Oxidized or reduced C423 of tetrameric PKM2 can suppress or enhance PUMA, respectively.

(A) PUMA mRNA expression and amounts of cleaved (*) caspase 3, Flag, and p53 were measured using qRT-PCR and immunoblots, respectively, in PKM-deficient A549 cells expressing WT or C423S-PKM2 treated with mitoxantrone and TEPP-46. 18S or actin was used as loading controls for qRT-PCR or immunoblots, respectively (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle, #P < 0.05 compared to mitoxantrone group). (B) Immunoprecipitation of p53 in PKM-deficient A549 cells expressing WT, C423S-, C423L-, C423D-, or C423E-PKM2 shown using immunoblots. (C) Immunoprecipitation of p53 in PKM-deficient A549 cells expressing C423S-PMK2 treated with vehicle or TEPP-46 shown using immunoblots. (D) Immunoprecipitation of p53 in PKM-deficient A549 cells expressing C423L-PMK2 treated with vehicle or TEPP-46 shown using immunoblots. (E) Immunoprecipitation of p53 in PKM-deficient A549 cells expressing C423E-PMK2 treated with vehicle or TEPP-46 shown using immunoblots. (F) PUMA mRNA expression was measured using qRT-PCR in PKM-deficient A549 cells expressing WT, C423L-, C423D-, or C423E-PKM2 treated with mitoxantrone and TEPP-46. The changes in PUMA mRNA expression with TEPP-46 are indicated with black and red or green arrows for WT and C423 mutants, respectively. 18S was used as a housekeeping gene (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle, #P < 0.05 compared to mitoxantrone). (G) PUMA mRNA expression was measured using qRT-PCR in PKM-deficient A549 cells expressing WT, C423D-, or C423E-PKM2 treated with bisantrene and TEPP-46. The changes in PUMA mRNA expression with TEPP-46 are indicated with black and blue arrows for the WT and C423 mutants, respectively. 18S was used as a housekeeping gene (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle, #P < 0.05 compared to bisantrene). (H) Pyruvate kinase activity was measured using a colorimetric assay that detects the oxidation of pyruvate in PKM-deficient A549 cells expressing WT, C423S-, C423L-, C423D-, or C423E-PKM2 treated with vehicle or TEPP-46 (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle). (I) Ribbon representation of PKM2 shows that C423 (CPK colors in white box) is localized on the periphery of the PKM2 monomer and near K433 (CPK colors in white circle), the p53 binding region of PKM2.

We speculated that if a mutation to S on residue 423 of PKM2 had completely mimicked a reduced cysteine, then it would have resulted in an increase in, rather than preventing the suppression of, mitoxantrone-mediated PUMA expression. However, S does not completely mimic a reduced form of cysteine, because it has a more negative value on the hydrophobicity scale (more polar) and a lower molecular weight (25). Thus, to further elucidate the effects of oxidized or reduced C423 of PKM2 on either p53 binding or p53 transcriptional activity, we generated additional C423 PKM2 mutants that have a similar structure, size, hydrophobicity (an index of polarity), molecular weight, and charge, either to reduced cysteine, which includes a mutation to leucine (L), or to oxidized cysteine (sulfonation), which includes mutations to aspartic acid (D) and E (see fig. S3A). We first assessed PKM2 binding to p53 and found that all C423 mutants and WT PKM2 were able to bind p53 (Fig. 3B), suggesting that C423 of PKM2 is not a critical residue for binding to p53. Furthermore, TEPP-46 increased the binding affinity between all of the C423 PKM2 mutants and p53 (Fig. 3, C to E, and fig. S3C), similar to what we observed with WT PKM2 (fig. S2D). Second, we investigated the function of reduced or oxidized C423 of PKM2 in the regulation of p53 transcriptional activity. TEPP-46 further increased mitoxantrone-mediated PUMA expression in C423L-PKM2, whereas it suppressed PUMA expression in WT PKM2 (Fig. 3F, compare red arrow to black arrow). Furthermore, although TEPP-46 did not further prevent mitoxantrone-mediated PUMA expression in C423D- and C423E-PKM2, compared to WT PKM2 (Fig. 3F, compare green arrows to black arrow), TEPP-46 was able to prevent bisantrene-mediated PUMA expression in C423D- and C423E-PKM2–expressing cells, compared to WT PKM2, where bisantrene-mediated PUMA expression was further increased (Fig. 3G, compare blue arrows to black arrow). Thus, our work provides evidence that the redox status of C423 on tetrameric PKM2 is critical for the differential regulation of p53 transcriptional activity.

Because stabilization of tetrameric PKM2 can also increase its enzymatic activity (19), we assessed whether PKM2 enzymatic activity is required for its regulation of p53 transcriptional activity. First, we reintroduced WT or C423 mutant PKM2 into PKM-deficient cells in the presence or absence of TEPP-46 and measured PKM2 enzymatic activity. We did not observe any major differences in PKM2 enzymatic activity between WT and C423S-, C423L-, C423D-, or C423E-PKM2 either at baseline or in the presence of TEPP-46, which increased PKM2 enzymatic activity similarly in all of the groups (Fig. 3H). Second, we generated a double PKM2 mutant at K305Q, which has diminished enzymatic activity (26), along with the C423 mutant forms of PKM2 that mimic either an oxidized or a reduced form of cysteine. We showed that similar to our single C423 mutant data in Fig. 3F, TEPP-46 further increased mitoxantrone-mediated PUMA expression in K305Q/C423L-PKM2, whereas it suppressed PUMA expression in K305Q-PKM2 (fig. S3D, compare red arrow to black arrow), which was comparable to WT PKM2 (Fig. 3F). In addition, TEPP-46 did not further prevent mitoxantrone-mediated PUMA expression in K305Q/C423D- and K305Q/C423E-PKM2 (fig. S3D), and once again, this is similar to what we observed in the C423D- and C423E-PKM2 mitoxantrone-treated cells (Fig. 3F). Overall, these data suggest that tetrameric PKM2-mediated regulation of p53 is not dependent on PKM2 enzymatic activity. Because C423 is localized on the periphery of PKM2, near the p53 binding site (Fig. 3I), it is plausible that oxidation of this residue alters the conformation of a tetrameric PKM2-p53 complex, resulting in inhibition of p53 transcriptional activity and apoptosis.

PKM2 is preferentially oxidized in the myocardium compared to the tumor

To assess whether PKM2 could be a potential therapeutic candidate in CIC, we first measured its relative expression and localization in cardiomyocytes. PKM2 is expressed in the myocardium but to a lesser degree than PKM1 (2729). Previous studies have reported the relative ratio of PKM1 to PKM2 mRNA at ~9 to 1, respectively (27, 29, 30). Using PKM2-deficient cardiomyocytes (see fig. S4A for methodology) and an antibody-based subtractive strategy, we confirmed a similar PKM1/PKM2 ratio at the protein level as well (fig. S4B). However, compared to PKM1, which is mainly localized in the cytoplasm, PKM2 had a nuclear-cytoplasmic localization in cardiomyocytes (Fig. 4A), suggesting that similar to cancer cells, PKM2 in cardiomyocytes may have nonenzymatic moonlighting functions in the nucleus, such as regulation of nuclear transcription factors (17).

Fig. 4 PKM2 is preferentially oxidized in the myocardium compared to the tumor.

(A) Immunofluorescence staining in isolated mouse cardiomyocytes showing PKM1 (green) and PKM2 (purple) localization using confocal microscopy. Nuclei are stained with DAPI in blue and indicated with arrowheads. Scale bar, 15 μm. Inset: Scale bar, 10 μm. (B) Transcription factor NRF2, along with the antioxidant proteins TRX and SFX, was measured in the myocardium and tumors of the same immunodeficient animals using immunoblots (means and SEM are shown, n = 3 different animals; *P < 0.05 compared to myocardium). Tubulin (tub.) was used as a loading control for immunoblots. (C) A schematic shows the oxidation and reduction cycle of glutathione, converting H2O2 to H2O in the process (top). GSH and GSSG were measured in myocardial and tumor tissues from immunodeficient animals (means and SEM are shown, n = 4 different animals; *P < 0.05 compared to myocardium). (D) ROS were measured by CellROX (green) in the myocardium and tumors of the same animals using confocal microscopy (means and SEM are shown, n = 4 different animals; *P < 0.05 compared to myocardium). The nuclear stain Hoechst is blue. Scale bar, 20 μm. AFU, arbitrary fluorescence units. (E) Global cysteine oxidation (as assessed by an antibody against cysteine sulfonation) was quantified in the myocardium and tumors of the same immunodeficient animals using immunoblots (means and SEM are shown on right, n = 3 different animals; *P < 0.05 compared to myocardium). Tubulin was used as a loading control for immunoblots. (F) PKM2 cysteine oxidation was measured in myocardium and tumors of the same immunodeficient animals by immunoprecipitation of PKM2 and an antibody against cysteine sulfonation using immunoblots (means and SEM are shown below, n = 3 different animals; *P < 0.05 compared to myocardium). Total immunoprecipitated PKM2 was used as a loading control. A long exposure of the inputs is shown in fig. S4H. (G) Mass spectrometry on immunoprecipitated PKM2 from the myocardium and tumor tissues identified C423 and C424 oxidation in the myocardium and reduced C423 and C424 in the tumor. m/z, mass/charge ratio.

To provide evidence that differential regulation of p53 transcriptional activity and apoptosis by tetrameric PKM2 could be therapeutically relevant in p53-induced cardiotoxicity, we next compared the oxidation state between the myocardium and the tumor in immunodeficient and immunocompetent models of A549 lung and E0771 mammary tumors, respectively. In both the immunodeficient (Fig. 4B) and immunocompetent (fig. S4C) models, the myocardium had lower expression of the antioxidant transcription factor nuclear factor (erythroid-derived 2)–like 2 (NRF2) (31), as well as the antioxidant proteins thioredoxin (TRX) and sulfiredoxin (SFX), compared to the tumor of the same animal. We also detected much lower amounts of reduced (GSH) and oxidized (GSSG) glutathione [a transcriptional target of NRF2 (32)] in the myocardium compared to the tumors (Fig. 4C and fig. S4D), suggesting that the tumors have a much higher buffering capacity for ROS than the myocardium. In keeping with these observations, ROS production was much higher in freshly excised myocardium compared to tumor biopsies of both tumor models (Fig. 4D and fig. S4E; see fig. S4F for background control signal). To study whether the increase in ROS correlated with protein cysteine oxidation, we probed myocardium and tumor lysates with an antibody against cysteine sulfonation and found an increase in its global levels in the myocardium compared to the tumor (Fig. 4E and fig. S4G). PKM2 cysteine oxidation was increased in the myocardium compared to the tumors of both immunodeficient and immunocompetent mice (Fig. 4F and fig. S4, H and I). Last, we detected oxidized C423 in the myocardium and reduced C423 in the tumor using high-resolution mass spectrometry (Fig. 4G). Therefore, we predicted that stabilization of tetrameric PKM2 with TEPP-46 would inhibit chemotherapy-induced p53 transcriptional activity and apoptosis in the myocardium but enhance them in the tumor.

TEPP-46 inhibits p53 transcriptional activity and apoptosis in cardiomyocytes but enhances p53 transcriptional activity and apoptosis in lung cancer cells

To assess whether TEPP-46 could inhibit chemotherapy-induced p53 transcriptional activity and apoptosis in myocardial cells but enhance them in tumor cells, we first isolated adult mouse cardiomyocytes and treated them with the cardiotoxic chemotherapy agent doxorubicin in the presence/absence of TEPP-46. We used a dose of doxorubicin that is similar to the concentrations detected in the serum of treated patients with cancer (33, 34) and has been previously shown to increase the p53 signaling pathway in human cardiomyocytes (35). At this dose of doxorubicin (150 nM), we did not observe any differences in PKM2 oxidation or oligomerization (fig. S5, A and B). TEPP-46 was able to inhibit p53 transcriptional activity (as assessed by Puma and p21 expression) and apoptosis [assessed by the apoptosis marker terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL)] in doxorubicin-treated cardiomyocytes (Fig. 5, A to C, and fig. S6, A and B). Furthermore, the induction of p53 transcriptional activity and apoptosis by doxorubicin was completely negated in p53-deficient cardiomyocytes, whereas the prevention of doxorubicin-mediated p53 transcriptional activity and apoptosis by TEPP-46 was completely negated in PKM2-deficient cardiomyocytes (Fig. 5, A to C, and fig. S6, A to D). Conversely, TEPP-46 was able to enhance p53 transcriptional activity and apoptosis in A549 and HTB177 lung cancer cells (Fig. 5, D to F, and fig. S6, E to J), and similar to cardiomyocytes, the induction of p53 transcriptional activity by doxorubicin was completely negated in p53-deficient cancer cells, whereas the enhancement of doxorubicin-mediated p53 transcriptional activity by TEPP-46 was completely negated in PKM2-deficient cancer cells (Fig. 5, D and E, and fig. S6, F and G). To confirm that C423 oxidation is important for TEPP-46–mediated inhibition of doxorubicin-induced p53 transcriptional activity in myocardial cells, we reintroduced either WT or C423S-PKM2 in PKM2-deficient cardiomyocytes. We found that TEPP-46 prevented doxorubicin-mediated Puma expression in WT but not C423S-PKM2–expressing cardiomyocytes (Fig. 5G). Conversely, in A549 cancer cells, we found that TEPP-46 was able to enhance doxorubicin-mediated PUMA expression in WT but inhibit doxorubicin-mediated PUMA expression in C423E-PKM2 mutant cells (Fig. 5G).

Fig. 5 TEPP-46 differentially regulates doxorubicin-induced p53 transcriptional activity and apoptosis in cardiomyocytes and lung cancer cells.

(A) Puma mRNA expression was measured using qRT-PCR in WT (p53+/+) and p53 knockout (p53−/−) isolated cardiomyocytes treated with doxorubicin and/or TEPP-46. 18S was used as a housekeeping gene (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle, #P < 0.05 compared to doxorubicin). (B) Puma mRNA expression was measured using qRT-PCR in WT (PKM2+/+) and PKM2 knockout (PKM2−/−) isolated cardiomyocytes treated with doxorubicin and/or TEPP-46. 18S was used as a housekeeping gene (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to PKM2+/+ vehicle, #P < 0.05 compared to doxorubicin, **P < 0.05 compared to PKM2−/− vehicle). (C) Apoptosis was measured in WT (p53+/+ and PKM2+/+) and p53 and PKM2 knockout (p53−/− and PKM2−/−) isolated cardiomyocytes treated with doxorubicin and/or TEPP-46 as assessed by TUNEL staining (green) using confocal microscopy. Representative images of the WT cells are shown on the left (a TUNEL-positive cell is indicated with white arrow), and quantified data are shown on the right. Nuclei were stained with DAPI in blue (n = 150 different cells from three fields; *P < 0.05 compared to vehicle, #P < 0.05 compared to doxorubicin). Scale bar, 20 μm. (D) PUMA mRNA expression was measured using qRT-PCR in A549 cells transfected with scrambled (Scr.; p53+/+) or p53 siRNA [p53 knockdown (KD)] treated with doxorubicin and/or TEPP-46. 18S was used as a housekeeping gene (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle, #P < 0.05 compared to doxorubicin). (E) PUMA mRNA expression was measured using qRT-PCR in A549 cells transfected with scrambled (PKM2+/+) or PKM2 siRNA (PKM2 knockdown) treated with doxorubicin and/or TEPP-46. 18S was used as a housekeeping gene (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle, #P < 0.05 compared to doxorubicin, **P < 0.05 compared to PKM2 knockdown control). (F) Apoptosis was measured in A549 and HTB177 cancer cells treated with doxorubicin and/or TEPP-46 as assessed by TUNEL staining (green) using confocal microscopy. Representative images are shown on the left, and quantified data are shown on the right. Nuclei are stained with DAPI in blue (means and SEM are shown, n = 150 different cells from three fields; *P < 0.05 compared to vehicle, #P < 0.05 compared to doxorubicin). Scale bars, 10 μm. (G) Puma mRNA expression was measured using qRT-PCR in PKM2-deficient isolated mouse cardiomyocytes expressing WT or C423S-PKM2 treated with doxorubicin and/or TEPP-46 (left). PUMA mRNA expression was measured using qRT-PCR in PKM-deficient A549 cells expressing WT or C423E-PKM2 treated with doxorubicin and/or TEPP-46 (right). 18S was used as a housekeeping gene (means and SEM are shown, n = 3 different experiments; *P < 0.05 compared to vehicle, #P < 0.05 compared to doxorubicin).

TEPP-46 prevents doxorubicin-induced cardiac dysfunction but enhances doxorubicin-induced lung tumor regression

We speculated that TEPP-46 would inhibit p53-mediated apoptosis in the heart, preserving cardiac function, while, at the same time, enhancing tumor apoptosis and tumor regression. To investigate this, we established an acute (2-week treatment protocol) and chronic (5-week treatment protocol) doxorubicin-induced cardiotoxicity model in mice with A549 and HTB177 lung tumors, respectively. Doxorubicin induced p53 in both the myocardium and the tumors (fig. S7, A and B), and it induced cardiac dysfunction, as assessed by decreased ejection fraction and fractional shortening compared to control-treated animals (Fig. 6, A and B, and fig. S8, A to C). TEPP-46 did not independently alter cardiac function but prevented doxorubicin-mediated cardiac dysfunction (Fig. 6, A to B, and fig. S8, B to C). Furthermore, TEPP-46 prevented a doxorubicin-mediated decrease in wall thickness, along with doxorubicin-induced interstitial fibrosis, resulting in an overall increase in survival (Fig. 6, C and D, and fig. S8, D to G). To confirm that the effects of TEPP-46 on cardiac function were mediated by PKM2, we treated WT- and cardiomyocyte-specific PKM2-deficient mice with doxorubicin in the chronic model of cardiotoxicity. TEPP-46 prevented doxorubicin-mediated cardiac dysfunction in WT but not in cardiomyocyte-specific PKM2-deficient mice (Fig. 6E and fig. S8H), confirming that the effects of TEPP-46 are mediated through PKM2 (19, 36).

Fig. 6 Stabilization of tetrameric PKM2 with TEPP-46 prevents doxorubicin-induced cardiac dysfunction but enhances doxorubicin-induced tumor regression in vivo.

(A) Cardiac function was measured using echocardiography as assessed by ejection fraction in mice with HTB177 tumors treated with doxorubicin and/or TEPP-46 in a 5-week treatment protocol in vivo (means and SEM are shown, n = 6 animals per group for control and TEPP-46, n = 10 animals per group for doxorubicin and doxorubicin + TEPP-46; *P < 0.05 compared to control-treated animals, #P < 0.05 compared to doxorubicin-treated animals). (B) Cardiac function was measured using echocardiography as assessed by fractional shortening in mice with HTB177 tumors treated with doxorubicin and/or TEPP-46 in a 5-week treatment protocol, in vivo (means and SEM are shown, n = 6 animals per group for control and TEPP-46, n = 10 animals per group for doxorubicin and doxorubicin + TEPP-46; *P < 0.05 compared to control-treated animals, #P < 0.05 compared to doxorubicin-treated animals). (C) Cardiac function was measured using echocardiography as assessed by diastolic left ventricular (LV) anterior wall thickness in mice with HTB177 tumors treated with doxorubicin and/or TEPP-46 in a 5-week treatment protocol, in vivo (means and SEM are shown, n = 6 animals per group for control and TEPP-46, n = 10 animals per group for doxorubicin and doxorubicin + TEPP-46; *P < 0.05 compared to control-treated animals, #P < 0.05 compared to doxorubicin-treated animals). (D) Interstitial cardiac fibrosis was assessed by Masson trichrome staining of heart slices in the 5-week treatment protocol. Means and SEM are shown on the left, and representative images of animals treated with doxorubicin or doxorubicin + TEPP-46 are shown on the right. Fibrotic tissue is stained in dark blue and is indicated with white arrowheads. Scale bar, 50 μm. (E) Cardiac function was measured using echocardiography as assessed by ejection fraction in WT and in cardiomyocyte-specific PKM2-deficient mice in a 5-week treatment protocol in vivo (means and SEM are shown, n = 6 for WT doxorubicin, n = 5 for WT doxorubicin + TEPP-46, n = 4 for cardiomyocyte-specific PKM2-deficient doxorubicin, n = 5 for cardiomyocyte-specific PKM2-deficient doxorubicin + TEPP-46 animals; *P < 0.05 compared to doxorubicin alone). (F) Percentage change from pretreatment of tumor volumes for the 5-week treatment protocol (HTB177) (means and SEM are shown, n = 6 control, n = 6 TEPP-46, n = 18 doxorubicin, n = 18 doxorubicin + TEPP-46; *P < 0.05 compared to control-treated animals, *#P < 0.05 compared to doxorubicin-treated animals; see fig. S8I for representative images). (G) Percentage change from pretreatment of tumor volumes for the 2-week treatment protocol (A549) (means and SEM are shown, n = 12 control, n = 6 TEPP-46, n = 14 doxorubicin, n = 14 doxorubicin + TEPP-46; *P < 0.05 compared to control-treated animals, *#P < 0.05 compared to doxorubicin-treated animals). (H) Apoptosis was measured in myocardial and tumor tissues from the 5-week (HTB177) and 2-week (A549) tumor models as assessed by TUNEL staining (green) and confocal microscopy. The nuclear stain DAPI is shown in blue (n = 15 to 17 different fields from five different animals per group; *P < 0.05 compared to control-treated animals, #P < 0.05 compared to doxorubicin-treated animals). White arrowheads indicate TUNEL-positive cells. Means and SEM are shown below (representative images above are for the 2-week myocardium and A549 tumor model; see fig. S8J for representative images of the 5-week myocardium and HTB177 tumor model). Scale bar, 40 μm. (I) Mechanism for differential regulation of p53 transcriptional activity and apoptosis by tetrameric PKM2. TEPP-46 binds at the monomer-monomer interface (and promotes tetrameric PKM2) on the opposite side of the p53 binding pocket via K433 of PKM2. Oxidation of C423 can change the conformation of p53-PKM2 interactions because it is in close proximity to K433. The oxidation state of the heart is higher than the tumor, resulting in preferential oxidation of PKM2 in the heart compared to the tumor. Tetrameric PKM2 can enhance chemotherapy-induced p53 transcriptional activity and apoptosis in the tumor, resulting in enhanced tumor regression, whereas tetrameric PKM2 oxidized on C423 in the heart can inhibit chemotherapy-induced p53 transcriptional activity and apoptosis, resulting in preserved cardiac function.

Next, we assessed the effects of TEPP-46 treatment on tumor progression in the same animals. Treatment with TEPP-46 decreased tumor growth compared to control-treated animals in both the acute and chronic p53-induced models (Fig. 6, F and G, and fig. S8I). Doxorubicin decreased tumor growth compared to control-treated animals, and combination treatment with TEPP-46 enhanced doxorubicin-mediated tumor regression (Fig. 6, F and G, and fig. S8I). To assess whether TEPP-46 limits doxorubicin-induced apoptosis in the myocardium while enhancing doxorubicin-induced apoptosis in the tumor, we probed for TUNEL and detected an induction of apoptosis with doxorubicin in both the heart and the tumor (compared to control-treated animals; Fig. 6H and fig. S8J). However, cotreatment with TEPP-46 prevented doxorubicin-induced apoptosis in the myocardium but enhanced doxorubicin-induced apoptosis in the tumor (Fig. 6H and fig. S8J). Furthermore, the prevention of doxorubicin-induced apoptosis in the myocardium by TEPP-46 was not observed in the cardiomyocyte-specific PKM2-deficient mice, further confirming the specificity of TEPP-46 to PKM2, in vivo (fig. S8K).

To validate the mechanism for TEPP-46–mediated differential regulation of p53-induced apoptosis between the heart (inhibition) and the tumor (induction), we first quantified PKM2, along with PKM2 binding to p53, in vivo. PKM2 expression remained consistent throughout the treatment protocol in both the acute and chronic studies (fig. S9A), and TEPP-46 increased the binding between PKM2 and p53 in the myocardium and the tumor (fig. S9, B and C). Furthermore, TEPP-46 completely prevented doxorubicin-induced mRNA expression of the p53 transcriptional target Puma in whole myocardial tissue and isolated cardiomyocytes (fig. S9, D and E), which was associated with a decrease in PUMA protein in the myocardium, but not in the tumors in both the acute and chronic doxorubicin-induced cardiotoxicity models (fig. S9, F and G).

Because doxorubicin increases mitochondrial ROS (37), it is plausible that doxorubicin could increase the oxidation of proteins (including PKM2) in the tumor. If this were to happen, then we would predict that TEPP-46 would inhibit, and not enhance, doxorubicin-induced tumor regression, as we observed in this study (Fig. 6, F and G, and fig. S8I). To address this, we measured global cysteine oxidation of proteins, along with PKM2 oxidation in the myocardium and tumor, from control- and doxorubicin-treated mice. Doxorubicin did not further increase global cysteine or PKM2 oxidation in either the myocardium or the tumor (fig. S10, A and B). These data suggest that although doxorubicin can increase ROS, it does not increase it to a degree that results in major changes to protein cysteine oxidation at physiologically relevant doses. Topoisomerase 2β (TOP2β) has previously been associated with doxorubicin-induced cardiotoxicity (37). In our study, we did not observe any changes in TOP2β expression in the myocardium between the different treatment groups (fig. S10C), suggesting that the main therapeutic benefit of TEPP-46 is mediated by inhibition of p53 and not other doxorubicin-mediated pathways.

DISCUSSION

Here, we show that stabilization of tetrameric PKM2 (with TEPP-46) in an oxidized environment, such as the myocardium, inhibits p53 transcriptional activity and apoptosis but enhances them in a more reduced environment, such as the tumor (Fig. 6I). Treatment of lung cancer tumors with TEPP-46 completely prevented doxorubicin-induced cardiac dysfunction but enhanced doxorubicin-induced tumor regression in the same animals. We propose that this tissue-specific differential regulation of apoptosis is because PKM2 is preferentially oxidized in the heart compared to the tumor.

We believe that the translational implications of our work are important for the prevention of doxorubicin-induced cardiac dysfunction, which remains untreatable. Although our findings are in keeping with a recent study that shows PKM2 can interact with p53 at its DNA response elements (20), a limitation of our work is that it does not characterize the precise mechanism by which the oxidized versus reduced form of PKM2 is able to change the conformation of p53, such that it inhibits or enhances, respectively, its transcriptional activity. Additional studies will be needed to elucidate potential differences in p53 binding conformation, interactions with other cofactors, or the ability to regulate the p53 transactivation domain by oxidized/reduced PKM2. Nevertheless, our results provide strong evidence that the redox status of C423, which is in close proximity to the p53 binding domain of PKM2 (K433), is critical for the differential regulation of p53 transcriptional activity. Because C423 is not important in the binding interactions between PKM2 and p53, this suggests that C423 of PKM2 could regulate a conformation of p53 such that p53 transcriptional activity is high when C423 is reduced or low when it is oxidized.

Because many cardiotoxic chemotherapeutic agents including trastuzumab, imatinib mesylate, and etoposide can induce p53 (3840), our work suggests that stabilizing tetrameric PKM2 in the myocardium could be beneficial to a large spectrum of CIC agents. Our work is in keeping with previous studies, showing that p53-deficient mice are resistant to anthracycline-induced cardiotoxicity (13) and that both p53- and PUMA-deficient mice are more resistant to afterload-induced myocardial apoptosis and heart failure (15, 41). Thus, the benefits of compounds that stabilize tetrameric PKM2 could extend beyond CIC to other types of heart failure. Similarly, our work also suggests that compounds similar to TEPP-46 may also be beneficial, independent of CIC, in increasing the therapeutic response to p53-activating cancer therapies, many of which are entering early-phase clinical trials (12). Quite often, the effects of chemotherapies on tumor regression are limited by CIC, and thus, we envision clinical trials where patients are pretreated or simultaneously treated with both tetrameric PKM2–stabilizing drugs (such as TEPP-46) and clinically approved chemotherapies.

MATERIALS AND METHODS

All animal studies were approved and performed under the University of Alberta Animal Care and Use Committee.

Study design

This study was designed to investigate the role of tetrameric PKM2 in the differential regulation of p53 transcriptional activity and to evaluate the preclinical efficacy of TEPP-46 (a compound that stabilizes tetrameric PKM2) in preventing doxorubicin-induced cardiac dysfunction. These objectives were addressed by (i) determining the oxidation type and site on PKM2 that is critical for the differential regulation of p53 transcriptional activity and apoptosis, (ii) evaluating the oxidation state in the myocardium and the tumor, (iii) evaluating the effects of TEPP-46 on doxorubicin-mediated p53 transcriptional activity and apoptosis in isolated cardiomyocytes and cancer cells, and (iv) evaluating the effects of TEPP-46 on cardiac function and tumor regression in lung tumors in mice treated with doxorubicin.

Animal sample sizes were calculated on the basis of a pilot study and previous animal studies with TEPP-46 (19). All data were included except if animals had a heart rate below 350 beats per minute (bpm) during echocardiography, which were about 10% of doxorubicin-treated mice. The end point of the study was the cumulative dose of doxorubicin (20 mg/kg), which was at the end of 2 or 5 weeks for the 10 or 4 mg/kg per week study, respectively. Mice in different cages were randomized to treatments, and investigators performing data collection and analysis were blinded to the groups.

Cre-LoxP heart-specific PKM2 knockout generation

C57BL/6J WT, α-myosin heavy chain (α-MHC)–Cre (stock no. 005657), and PKM2 flox (stock no. 024048) mice having LoxP sites flanking exon 10 of the PKM gene were purchased from the Jackson Laboratory. To generate PKM2 cardiac−/− mice, α-MHC–Cre transgenic mice expressing tamoxifen-inducible Cre in cardiomyocytes were bred with PKM2-flox mice. Genotypes of mouse offspring were confirmed by PCR of Cre as amplification of a 100–base pair (bp) fragment or presence of a floxed PKM2 or WT PKM2 gene as amplification of a 680- or 578-bp fragment, respectively (primer sequences 19311 and 19312 from the Jackson Laboratories). Cre-induced inactivation of the PKM2 gene was carried out via six intraperitoneal injections of tamoxifen (50 mg/kg) spread over 8 days in mice starting at 6 to 7 weeks of age (42). All mice were allowed 5 weeks washout after tamoxifen administration before experimentation, due to the transient cardiomyopathy phenotype induced by cardiac-specific Cre expression (43). Male WT- or cardiomyocyte-specific PKM2-deficient mice were randomly assigned to sham treatment or treatment with doxorubicin (10 mg/kg; Teva Standard) given as an intraperitoneal injection weekly for 5 weeks and/or TEPP-46 (100 mg/kg; MedChemExpress; dissolved in 0.5% carboxymethylcellulose and 0.1% Tween 80) given daily by oral gavage.

In vivo tumorigenicity models

Male nu/nu (6 weeks old; Charles River Laboratories) or female C57BL/6 (the Jackson Laboratory) mice were subcutaneously injected with 3 × 106 A549, EO771, or HTB-177 cells suspended in phosphate-buffered saline and matrigel (Corning) in a 1:1 ratio in the right flank. Once tumors became visible (~2 weeks), animals were randomly assigned to sham treatment or treatment with doxorubicin (10 mg/kg; Teva Standard) given as an intraperitoneal injection weekly for 2 or 5 weeks and/or TEPP-46 (100 mg/kg; MedChemExpress; dissolved in 0.5% carboxymethylcellulose and 0.1% Tween 80) given daily by oral gavage for the duration of the experiment. Animal numbers were chosen on the basis of previously published data (19).

Echocardiography

Mice were anesthetized with isoflurane and were maintained at a heart rate of 350 to 450 bpm. Echocardiography was performed as previously described (44) using the Vevo 770 High-Resolution Imaging System (VisualSonics). Both B- and M-mode images were acquired, and measurements for left ventricular size and function were taken in the parasternal short axis view using the papillary muscles as landmarks for the mid-ventricular view. Ejection fraction, fractional shortening, and wall thickness were calculated from M-mode. Images with heart rate of <350 bpm were excluded from the analysis.

Statistics

Statistical analysis was performed on SPSS version 24.0 (Somers). Values are expressed as means ± SEM. Wilcoxon test was used for Fig. 2D as a nonparametric paired samples test. Log-rank (Mantel-Cox) test was used for fig. S8G to measure distribution of data. A nonparametric Levene’s test was used to assess equality of variances between the samples. All subsequent nonparametric tests were only performed if the equality of variances was met. Kruskal-Wallis test was used to compare two or more independent samples, and the Mann-Whitney U test was used for comparisons between two groups. Significance was considered at P < 0.05.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/11/478/eaau8866/DC1

Materials and Methods

Fig. S1. WT, but not K433E PKM2, can interact with p53, and PKM1 and K433E PKM2 do not reverse the increase in p21 and PUMA mRNA in PKM-deficient cells.

Fig. S2. Oxidation of tetrameric PKM2 inhibits p53 transcriptional activity.

Fig. S3. C423 of tetrameric PKM2 is the critical residue required for differential regulation of p53 transcriptional activity.

Fig. S4. PKM2 is localized in the cytoplasm and nucleus of cardiomyocytes, and the myocardium is more oxidized than the tumor.

Fig. S5. Doxorubicin does not increase PKM2 oxidation or oligomerization.

Fig. S6. TEPP-46 differentially regulates doxorubicin-induced p53 transcriptional activity and apoptosis in cardiomyocytes and lung cancer cells.

Fig. S7. Doxorubicin induces p53 in the myocardium and tumor, and PKM2 is present in both myocardium and tumor tissues.

Fig. S8. TEPP-46 prevents chemotherapy-induced cardiac dysfunction but increases chemotherapy-mediated tumor regression.

Fig. S9. TEPP-46 prevents doxorubicin-induced Puma increase in the myocardium but not the tumor in vivo.

Fig. S10. Doxorubicin does not further increase global protein cysteine sulfonation, PKM2 sulfonation, or TOP2β expression.

Table S1. Primer sequences for site-directed mutagenesis experiments.

REFERENCES AND NOTES

Acknowledgments: We would like to thank the Alberta Proteomics and Mass Spectrometry Facility at the University of Alberta for mass spectrometry acquisition. Funding: B.S. is supported by a graduate scholarship from Alberta Innovates. J.R.U. is supported by a grant-in-aid from the Heart and Stroke Foundation of Canada. G.S. is supported by an Alberta Innovates Translational Health Chair in cardio-oncology and a National and Alberta New Investigator award from the Heart and Stroke Foundation of Canada, along with grants from the Canadian Institutes of Health Research and Mazankowski Alberta Heart Institute/Heart and Stroke Foundation of Alberta. Author contributions: B.S. performed all experiments, except for mass spectrometry, cardiomyocyte isolation, generation and characterization of cardiomyocyte-specific PKM2-deficient mice, and echocardiography, analyzed and interpreted data, and cowrote the manuscript. V.G. performed echocardiography. Y.Z. isolated and cultured adult cardiomyocytes. A.K., along with B.S. and G.S., generated all in vivo tumor models. A.E.B., along with B.S., performed tissue immunoblots. K.G. and J.R.U. developed and characterized the cardiomyocyte-specific PKM2-deficient mice. G.S. generated the hypothesis, designed, funded, and supervised the study, performed experiments, interpreted data, and cowrote the manuscript. All authors edited the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article