KLF15 Regulates the Circadian Susceptibility to Ischemia Reperfusion Injury in the Heart

The onset and the injury of myocardial infarctions have a circadian rhythmicity with the peak in the early morning in humans. The molecular mechanism, however, remains poorly understood. Transcription factor Kruppel-like factor 15 (KLF15) regulates 75% of the oscillatory transcripts in the heart and has a peak expression at ZT14, the beginning of the active phase (Figure 1A).¹ KLF15 expression is significantly reduced in the heart tissue of patients with cardiomyopathies (the study was approved by an institutional review committee, Figure 1B).²

Figure.

KLF15 regulates circadian I/R susceptibility of the heart.

A, Mouse cardiac KLF15 expression during a 24-hour day by immunoblot. ZT, zeitgeber time. ZT 22 was set to 1 (n = 3 per time point, *: p< 0.05, **: p< 0.01, ***: p<0.001 vs. ZT22, one-way ANOVA with Bonferroni correction). Shading indicates the light and dark periods under 12:12 light-dark conditions. B, qRT-PCR of Klf15 in human heart samples from patients with ischemic cardiomyopathy (ICM) and controls. (*: p< 0.05, 2-sided Student’s t-test). C, Top, Experiment design. Bottom left, Representative TTC staining. Bottom right, Percent of infarct over area at risk. (n=8–13, *: p<0.05 vs. controls at ZT2; #: p<0.05 vs. controls at ZT 14, two-way ANOVA with Bonferroni correction correction). D, all samples are collected at ZT14 except top left panel. Top left, MnSOD activity in cKLF15 KO mice and control mice hearts. (n=6–7, *: p<0.005 vs. control at ZT2, **: p<0.0001 vs. control at ZT2, #: p<0.005 vs control at ZT14, two-way ANOVA with Bonferroni correction). Top right, MnSOD and MnSOD^K122 acetylation level measured by immunoblot. (n=4, *: p<0.001, 2-tailed Student’s t test, Holm-Sidak correction). Middle left, Cardiac mitochondrial total protein acetylation of cKLF15 KO mice and littermate control mice. (n=3–4). Middle right, SIRT3 protein level determined by immunoblot blot. Bottom, Cardiac total NAD, NAD⁺ level and NAD⁺/NADH ratio from cKLF15 KO and littermate control mice. (n=6–7, *: p<0.05 vs. litter mate controls, 2-tailed Student’s t test, Holm-Sidak correction). E, Top, Experiment design. Middle left, Representative TTC staining. Middle right, Percent of infarct over area at risk. (n=6, *: p<0.01vs. littermate controls without NMN, #: p<0.05 vs. cKLF15 KO mice without NMN, two-way ANOVA with Bonferroni correction). Bottom, cardiac total NAD, NAD⁺ level and NAD⁺/NADH ratio from cKlf15 KO and littermate control mice pretreated with NMN/PBS prior to I/R injury at ZT14 (n=4, *; p<0.05 vs. litter mate control without NMN, #: p<0.01 vs. cKlf15 KO mice without NMN; 2-tailed Student’s t-test, Holm-Sidak correction). F, Top, Total NAD pool, NAD⁺, and the NAD⁺/NADH ratio in cKLF15 KO mice and littermate control hearts. (n = 3–4 per time point, *: p<0.05, **: p<0.01, ***: p<0.001, 2-tailed Student’s t test, Holm-Sidak correction). Middle left, Experiment design and representative TTC staining. Middle right, Percent of infarct over area at risk (bottom right). (n=6–7, *: p<0.01vs. mice without NMN, 2-tailed Student’s t test, Holm-Sidak correction). Bottom, cardiac total NAD, NAD⁺ level and NAD⁺/NADH ratio from control mice pretreated with NMN/PBS prior to I/R injury at ZT2 (n=3, *: p<0.05 vs. mice without NMN, 2-tailed Student’s t-test, Holm-Sidak correction). G, Top, cardiac Nampt RNA level in cKlf15 KO mice and littermate controls in 24hrs. (n = 3–6; *: p<0.05, two-way ANOVA with Bonferroni correction). Bottom, cardiac NAMPT protein level determined by immunoblot blot. (n=4, **: p<0.01, ***: p<0.001, ****: p<0.0001, two-way ANOVA with Bonferroni correction). H, Luciferase reporter assays in 3T3 cells. Genomic locus of mouse Nampt promoter region was shown adapted from the UCSC genome browser. ENCODE data for H3K27ac in 8 weeks mouse heart is shown to indicate 2 enhancer regions. Putative KLF15 binding sites were identified based on consensus sequence, shown in vertical lines. A series of promoter sequence, and a 229bp fragment with a 21bp deletion from the 250bp fragment (-[ ]-) were cloned into pGL4.23 vector. 3T3 cells with shRNA knockdown of scrambled RNA (sh-Sc) or KLF15 (sh-Klf15) were transiently co-transfected with the luciferase vector for 48 hours before luminescence was measured. (n=4, *: p< 0.05 vs. sh-Sc, #: p<0.05 vs. sh-Sc transfected with empty vector, two-way ANOVA with Bonferroni correction). I, Working model.

To investigate if KLF15 affects myocardial injury in a diurnal fashion, we subjected cardiomyocyte-specific KLF15 knockout mice (cKLF15 KO) and their litter mate controls (Klf15^flox/flox) to ischemia reperfusion (I/R) injury (all procedures in this study were in accordance with institutional guidelines). Similar to previously reported³, the control mice showed a smaller infarct at ZT14 than ZT2. However, the cKLF15 KO mice had bigger infarcts at both times, similar to the size of the control group at ZT2 when KLF15 troughs, suggesting that myocardial KLF15 is critical for the reduced cardiac injury at ZT14 (Figure 1C).

Reperfusion injury is characterized by a reactive oxygen species burst from the mitochondria electron transport chain. We next examined the activity of MnSOD, the mitochondrial superoxide dismutase isoform. We found that cardiac MnSOD activity is higher at ZT14 than ZT2 in the control hearts, and it is further reduced in the cKLF15 KO mice heart at both times, which is associated with hyperacetylation at MnSOD^K122 and unchanged protein expression levels. A global hyperacetylation of mitochondrial proteins is also observed in the cKLF15 KO mice heart. While the main mitochondria deacetylase SIRT3 expression remained unchanged, we found its co-enzyme NAD⁺ was reduced in the KLF15 deficient heart at ZT14. These results suggest that the absence of KLF15 is associated with NAD⁺ deficiency, which then leads to SIRT3 dysfunction and hyperacetylation of mitochondria proteins including MnSOD and subsequently leads to reduced activity of MnSOD and likely other mitochondrial proteins (Figure 1D).

NAD⁺ was reported to be beneficial during cardiac I/R.⁴ Given the pronounced reduction in NAD⁺ in the heart of cKLF15 KO, we attempted to rescue by supplementing the cell and mitochondria permeable NAD⁺ precursor, nicotinamide mononucleotide (NMN). A single dose of NMN 500mg/kg was given 30 minutes before ischemia intraperitoneally, which allowed NAD⁺ and NAD⁺/NADH ratio in the cKLF15 KO to be corrected to close to control levels. There was a significant reduction of the infarct in the cKLF15 KO mice post NMN supplementation indicating the exaggerated I/R injury at ZT14 in cKlf15 KO mice was in fact due to NAD⁺ deficiency (Figure 1E).

We noted that the infarction of the control mice was not significantly improved, which we suspected may be due to the circadian variation of NAD⁺ level in the heart. NAD⁺, total NAD⁺ pool and the NAD⁺/NADH ratio all have a diurnal oscillation in the control mice. As NAD⁺ level troughs during the dark to light transition for the control mice, we hypothesized that NMN administration at this time (ZT2) may be more beneficial. We then performed an NMN rescue I/R experiment in the control mice at ZT2. Indeed, we observed a reduced infarct size accompanying improved NAD⁺ levels (Figure 1F).

Interestingly, the rise of NAD⁺ during the dark phase seen in the control mice hearts is completely abolished in the cKLF15 KO mice hearts (Figure 1F). Nampt has been shown to be the determinant factor in cardiac NAD⁺ level.⁵ Nampt messenger RNA shows a circadian rhythmicity and rises quickly after the initial KLF15 rise at ZT2. Remarkably, this oscillation is completely abolished in the heart of cKlf15 KO mice, suggesting that although the baseline expression of Nampt is independent of KLF15, the circadian rhythmicity and specifically the peak expression at ZT6 is critically dependent on the presence of KLF15 in the heart. NAMPT protein expression also exhibits a similar trend (Figure 1G).

Because KLF15 regulates Nampt expression at the transcript level, we next examined the promoter/enhancer region of Nampt using luciferase assay in NIH-3T3 cells co-infected with adenoviral vector expressing Klf15 knockdown shRNA (sh-Klf15) or control shRNA (sh-sc). With serial deletion constructs, we were able to confirm a 21bp region containing two consecutive KLF15 binding sites is the KLF15 regulated enhancer site that controls the circadian rhythmicity of Nampt expression in the heart (Figure 1H).

In summary, we show that oscillatory transcription factor KLF15 contributes to the susceptibility of myocardium to I/R injury in a time-of-the-day specific fashion (Figure 1I). KLF15 is reduced in chronic cardiac diseases, similar to KLF15 deficient mice, which may lead to NAD⁺ deficiency specifically during the sleep to active transition and may increase the susceptibility of the myocardium to I/R injury. Our study suggests that NAD⁺ supplementation during a critical time window maybe particularly beneficial for patients with chronic heart diseases.