Hsp20 Interacting with Phosphorylated Akt Reduces Doxorubicin-Triggered Oxidative Stress and Cardiotoxicity

Effects of DOX on the Expression of Major Hsps in the Murine Heart

To assess the regulatory roles of cardiac Hsps in response to DOX treatment, we examined the expression profiles of the six major Hsps (Hsp90, Hsp70, Hsp60, Hsp27, αB-crystallin and Hsp20) at different time points, following DOX ip injection. Western blots and quantitative results (Figure 1) indicated that alterations in the levels of Hsp70 and Hsp60 exhibited a biphasic response: a) they greatly increased at 1 hour after DOX injection, and returned to basal levels by 2 hours; b) they increased again at 4-12 hours, and significantly decreased below basal levels at 2-3 days after DOX treatment. There was no significant change in Hsp90 expression, while DOX increased Hsp27 content by 2 fold at 30 min, and this increase was maintained up to 3 days. The levels of αB-crystallin were transiently increased at 30 min to 1 hour, and then returned to basal. While a similar transient increase was also observed with Hsp20 expression, its levels decreased by 40% at 2-3 days after DOX injection. These results demonstrate that the expression pattern of various Hsps is altered in the heart upon DOX insult, suggesting that these Hsps may be involved in DOX-induced cardiomyopathy via different mechanism(s).

DOX-Induced Cardiomyocyte Death and Apoptosis is Suppressed by Increased Hsp20 Levels

Several studies have shown that some Hsps can act as negative regulators of DOX-triggered apoptotic and necrotic cell death, such as Hsp10 and Hsp60, which have been found to modulate DOX-induced mitochondrial apoptosis signaling in neonatal cardiomyocytes.¹⁶ However, the underlying mechanisms are still remained to be clarified. To investigate whether increased Hsp20 levels could have an inhibitory effect on DOX induced cardiomyocyte death, we first infected H9c2 cells with Ad.Hsp20 or Ad.GFP for 24 h and subsequently subjected them to DOX treatment. After 24 h of treatment with DOX at 0.5μM, 40% of the Ad.GFP-infected cells were not viable as observed by phase-contrast microscopy (Figure 2A), and measured by the MTS assay (Figure 2B). However, DOX-induced cell death was markedly diminished in the Ad.Hsp20-infected group (Figure 2A and B). The cytoprotective effect of Hsp20 was also observed in cells treated with different concentrations of DOX or infected with different concentrations of Ad.Hsp20, which was dependent on the level of Hsp20 overexpression (Figure 2C and D), suggesting that Hsp20 acts directly on cardiomyocytes to inhibit DOX-induced cell death.

To further confirm that Hsp20 acts as a survival factor in cardiomyocytes, ventricular myocytes were isolated from adult rat hearts, and infected with Ad.GFP or Ad.Hsp20. The infection efficiency reached more than 95% after 24 h (Figure 2E). These infected-cells were subsequently subjected to 20 μM DOX treatment for 24 h. Both MTS and DNA fragmentation assays showed that ectopic overexpression of Hsp20 protected adult rat cardiomyocytes against DOX-triggered cell death and apoptosis (Figure 2E).

Decreased DOX-Induced Cardiomyocyte Injury and Apoptosis in Hsp20-TG Hearts

To determine whether Hsp20 overexpression protects cardiomyocytes against DOX-induced apoptosis in vivo, TG mice overexpressing Hsp20 under the control of the cardiac-specific α-MHC promoter were analyzed. TG hearts showed a ten-fold increase in total (transgenic and endogenous) Hsp20 content, compared with WT mouse hearts.¹³ Eight-week-old WT and TG mice were subjected to one intraperitoneal (ip) injection of 20 mg/kg of DOX or saline as a control. At 4 days after the injection of DOX, the hearts of DOX-treated WT mice (Figure 3A), but not those of DOX-treated TG mice (Figure 3B), displayed focal cytoplasmic vacuolization, a hallmark of cell injury, which is consistent with previous reports.^{17, 18} The percentage of TUNEL-positive nuclei in the DOX-treated WT hearts was 2.5-fold higher than in DOX-treated TG hearts (Figure 3C). Accordingly, DNA fragmentation, measured by an ELISA method, was reduced by 30% in DOX-treated TG hearts, compared with controls (Figure 3D). These results indicate that overexpression of Hsp20 in vivo significantly decreased DOX-induced cardiac injury and apoptosis.

Preserved Cardiac Function in Hsp20-TG Mice after Acute DOX Administration

Recent studies have shown that very low levels of myocyte apoptosis (23 myocytes per 10⁵ nuclei of animal hearts; 80-250 myocytes per 10⁵ nuclei of human hearts) significantly affect cardiac function and survival.¹⁹ Given the marked disparity of cell injury and apoptosis in DOX-treated WT and Hsp20-TG hearts, we examined contractile function and animal survival in a separate series of experiments. 4 days after the injection of 20 mg/kg of DOX or saline, WT and TG mouse hearts were subjected to functional measurement by the Langendorff mode. As shown in Figure 4 A-D, DOX treatment caused a significant decrease in the rates of contraction (+dP/dt) and relaxation (−dP/dt) in WT hearts, accompanied by a significantly increased left ventricular end-diastolic pressure (LVEDP) and reduced left ventricular developed pressure (LVDP). However, these functional parameters showed no significant difference in TG hearts (Figure 4A-D). In addition, while 6 out of 11 WT mice died within 12 days after treatment with DOX, most TG mice (12 out of 14) survived for up to 12 days after treatment (Figure 4E, P<0.05, vs WT). These findings indicate that overexpression of Hsp20 in the heart reduced the extent of DOX-induced acute heart failure and prolonged survival.

Improved Cardiac Function and Survival in Hsp20-TG Mice after Chronic DOX Administration

Since patients typically receive lower levels of DOX given over many weeks, ^{2, 3} we further investigated whether cardiac-specific overexpression of Hsp20 could improve myocardial function and prolong survival in a chronic model of DOX-induced cardiotoxicity. To develop such a chronic model, an initial DOX dose of 4 mg/kg was selected and mice were injected weekly for 5 weeks. The survival curve of the treated groups indicated that there was no significant difference between WTs and TGs; although all the treated WT mice died, whereas 3 out of 9 treated TG mice survived during the 10 weeks follow-up after the last DOX injection (Figure 5A). However, in an additional chronic model, mice received 3 mg/kg weekly for 6 weeks, and we observed that all the treated TG mice survived during the 7 weeks follow-up after the last DOX administration; whereas 5 out of 12 treated WT mice died (Figure 5B, P<0.05 vs WT). To determine the alteration of myocardial function in this chronic model, one week after the last administration of DOX (3 mg/kg weekly for 6 weeks), we measured myocardial contractility by the Langendorff preparation to avoid the in vivo confounding effects, such as loading conditions and neurohormonal factors. As shown in Figure 5C, the rates of contraction (+dP/dt) and relaxation (−dP/dt) were significantly improved, accompanied by a reduced LVEDP and an increased LVDP in DOX-treated TG hearts, compared with DOX-treated WT hearts.

Akt-BAD Signaling Pathway is Preserved in DOX–Treated Hsp20-TG Hearts

To elucidate the possible mechanism(s) of Hsp20 cardioprotection against DOX, we first assessed other major Hsps' expression levels in acute DOX-treated TG hearts and controls. Western blots showed that overexpression of Hsp20 did not alter the expression levels of other major Hsps (Hsp90, Hsp70, Hsp60, Hsp27, and αB-crystallin) either in DOX-treated or saline-treated hearts, compared with WT hearts (Online Figure I), suggesting that Hsp20 overexpression does not result in compensatory adaptation by other Hsps, and clearly indicates that the sole increase in Hsp20 levels can lead to protection against DOX-induced cardiotoxicity.

A number of studies have shown that two classes of cell growth and survival/death signaling pathways, the PI3K/Akt and MAPKs pathways, play important roles in regulating cell survival. ^{20, 21} Our results showed that there were no differences in activation of MAPK signaling between WT hearts and TGs after acute DOX administration (Online Figure II). However, DOX-treated WT hearts revealed significantly decreased activation of Akt (generally referred to as Akt1 hereafter), as determined by immunoblotting for phosphorylation of either Thr308 or Ser473. Importantly, phosphorylation of Akt was preserved in DOX-treated TG hearts (Figure 6A), which was further supported by in vitro kinase activity assays, using GSK-3α/β as a substrate (Figure 6B). Surprisingly, we found that this immuno-complex of pSer 473-Akt contained Hsp20 (Figure 6B). Furthermore, reciprocal co-immunoprecipitation with an Hsp20 antibody showed an increased association between Hsp20 and phosphorylated Akt in DOX-treated Hsp20-TG hearts (Figure 6C). To further examine whether Hsp20 interacts with non-phosphorylated isoform of Akt, we co-infected H9c2 cells with a nonphosphorylatable form of Akt1-AAA (Ad.dnAkt) and Ad.Hsp20. While Ad.Hsp20/Ad.dnAkt-cell lysates revealed a significant increase in total Akt level (the most is exogenous non-phosphorylated Akt1), immunoprecipitation experiments with the Hsp20 antibody showed that Hsp20-associated Akt levels were similar to those in the Ad.GFP control (Figure 6D), suggesting that Hsp20 may selectively interact with phosphorylated Akt. In addition, co-immunoprecipitation results indicated that Hsp20 was associated with not only Akt1, but also Akt2 (Online Figure III) in the murine heart.

Given the marked disparity of Akt activity in DOX-treated hearts, we examined its downstream targets to assess their potential role in DOX-induced cardiotoxicity. Akt has been shown to promote cell survival via its ability to phosphorylate BAD at Ser136.^{22, 23} Accordingly, pS136-BAD was markedly reduced in DOX-treated WT hearts, but not in DOX-treated TG hearts (Figure 6E). However, another critical downstream element of the Akt cell survival pathway, GSK-3β phosphorylation (Ser 21), was unaltered in either WT or TG hearts after DOX treatment (data not shown). These results suggest that pBAD alteration could be attributed to the anti-apoptotic effects of Hsp20. Also, an active form of caspase 3, cleaved caspase-3, was 1.8-fold higher in WT than TG upon DOX treatment (Figure 6E). Taken together, these findings indicate that Hsp20 interacts with phosphorylated Akt, and maintains its phosphorylated state, leading to preserved Bad phosphorylation and inhibition of caspase-3 cleavage, which consequently results in attenuation of DOX-triggered cardiomyocyte apoptosis.

It has been documented that inactivation/dephosphorylation of Akt can be regulated by Ser/Thr phosphatases (PP1 and PP2).^23-25 As such, we next examined whether the levels of phosphatase (PP1, PP2A, or PP2B) expression were altered in the DOX-treated hearts. As shown in Figure 6F, protein phosphatase 1 (PP1) was significantly upregulated in both DOX-treated WT and TG hearts, while PP2a and PP2B were not altered in either condition, suggesting that the Hsp20-p-Akt complex may block the action of PP1 on Akt dephosphorylation.

Maintenance of Akt Activation Is Required in Protection of Hsp20 against DOX-Induced Cardiomyocyte Death

To further confirm our in vivo findings on involvement of phosphorylated Akt in the Hsp20 cardioprotection against DOX, we inhibited activation of Akt and MAPKs or their upstream kinases prior to DOX addition in either Ad.GFP- or Ad.Hsp20-infected adult rat cardiomyocytes. Quantitation of cell survival with MTS revealed that pre-treatment with either the p38 inhibitor SB203580, MEK1/2 inhibitor PD98059, or JNK inhibitor had no apparent effect on protection of Hsp20 against DOX-induced cell death (Figure 7A). However, upon addition of the PI3K inhibitors Ly29402 and wortmannin, which inhibit phosphorylation/activation of its substrate Akt, the protective action of Hsp20 was eliminated in DOX-treated cardiomyocytes (Figure 7B). Furthermore, infection with a dominant-negative form of Akt (dnAkt) also offset the protection of Hsp20 against DOX-triggered cardiomyocyte death (Figure 7C). Taken together, these results suggest that Hsp20 protects against DOX-induced cardiomyocyte death mainly via the Akt-dependent signaling pathway.

Attenuation of Oxidative Stress in the Hsp20-TG Hearts Following DOX Treatment

DOX-induced cardiotoxicity has previously been suggested to be attributable mostly to cardiomyocyte damage caused by increased oxidative stress.^{26, 27} To determine if increased Hsp20 expression might be accompanied by the induction of antioxidant pathways, which detoxify reactive oxygen species (ROS), we next measured the activities of antioxidant enzymes and ROS levels in heart homogenates. The enzymatic activities of both superoxide dismutase (SOD) and glutathione peroxidase (GPx) were significantly reduced in DOX-treated hearts, compared with saline-treated samples. However, these decreases were less pronounced in TG hearts, relative to WTs (Figure 8A and B). Similarly, the ROS levels were markedly increased in WTs after DOX treatment, but overexpression of Hsp20 attenuated this increase (Figure 8C). Furthermore, increased Hsp20 expression could effectively suppress DCF fluorescence in DOX-stressed cardiomyocytes (H9c2) in a dose-dependent fashion (Figure 8D), consistent with its anti-apoptotic effects (Figure 2D).