Volume 95, Issue 5, Pages 1120-1137 (May 2019) β-hydroxybutyrate attenuates renal ischemia-reperfusion injury through its anti- pyroptotic effects  Takaya Tajima, Ayumi Yoshifuji, Ayumi Matsui, Tomoaki Itoh, Kiyotaka Uchiyama, Takeshi Kanda, Hirobumi Tokuyama, Shu Wakino, Hiroshi Itoh  Kidney International  Volume 95, Issue 5, Pages 1120-1137 (May 2019) DOI: 10.1016/j.kint.2018.11.034 Copyright © 2019 International Society of Nephrology Terms and Conditions

Kidney International 2019 95, 1120-1137DOI: (10.1016/j.kint.2018.11.034) Copyright © 2019 International Society of Nephrology Terms and Conditions

Figure 1 The effects of β–hydroxybutyrate (β-OHB) on renal function and histology. (a) β-OHB plasma levels were measured 24 hours before and after sham or ischemia-reperfusion (IR) operation. (b,c) Blood urea nitrogen levels (b) and serum levels of creatinine (c) in each mice group; phosphate-buffered saline–treated sham-operated mice, β-OHB-treated sham (sham+β-OHB), phosphate-buffered saline–treated IR-operated mice, and β-OHB-treated IR (IR+β-OHB). (d) Representative kidney section stained for hematoxylin and eosin in each mice group. Arrowheads indicate inflammatory cells or cell debris. Bar graph in the lower panel represents semiquantitative analysis of tubule interstitial injury. (e) Representative kidney section stained with periodic acid–Schiff in each mice group. Arrowheads indicate brush border. Bar graph in the lower panel represents semiquantitative analysis of tubular injury. (d,e) Original magnification ×200. Insets: increased magnification ×500. Bars = 50 μm. Insets are more magnified view of tubular cells. Inset bars = 20 μm. ##P < 0.01 versus sham, **P < 0.01 versus IR, *P < 0.05 versus IR. n = 6 for sham and sham+β-OHB, n = 8 for IR and IR+β-OHB. Bar represents mean ± SEM. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org. Kidney International 2019 95, 1120-1137DOI: (10.1016/j.kint.2018.11.034) Copyright © 2019 International Society of Nephrology Terms and Conditions

Figure 2 The effects of 2-day fasting and β-hydroxybutyrate (β-OHB) on renal function in mice with ischemia-reperfusion (IR) injury. (a–d) The effects of 48-hour fasting on IR injury. Body weights (a) and β-OHB plasma levels (b) were measured 48 hours before and at the time of the sham or IR operation. Blood urea nitrogen levels (c) and serum levels of creatinine (d) were measured 24 hours after the sham or IR operation. ##P < 0.01 versus nonfasted sham-operated mice (sham+nonfasted), #P < 0.05 versus sham+nonfasted, **P < 0.01 versus nonfasted IR-operated mice (IR+nonfasted), and *P < 0.05 versus IR+nonfasted. n = 3 for sham+nonfasted and sham+fasted; n = 4 for IR+nonfasted and IR+fasted. (e–h) The combined effects of β-OHB treatment and fasting on IR injury. Body weights (e) and β-OHB plasma levels (f) were measured 48 hours before and at the time of the sham or IR operation. ##P < 0.01 versus phosphate-buffered saline–treated nonfasted mice with IR operation (IR+nonfasted), #P < 0.05 versus IR+nonfasted, and **P < 0.01 versus β-OHB–treated nonfasted mice with IR operation (IR+nonfasted+β-OHB). Blood urea nitrogen levels (g) and serum levels of creatinine (h) were measured 24 hours after the sham or IR operation. ##P < 0.01 versus phosphate-buffered saline–treated nonfasted mice with sham operation (sham+nonfasted), **P < 0.01 versus IR+nonfasted, *P < 0.05 versus IR+nonfasted, $P < 0.05 versus IR+nonfasted+β-OHB, ¶¶P < 0.01 versus phosphate-buffered saline–treated fasted mice with IR operation (IR+fasted), and ¶P < 0.05 versus IR+fasted. n = 6 for each mouse group. Bar represents mean ± SEM. Kidney International 2019 95, 1120-1137DOI: (10.1016/j.kint.2018.11.034) Copyright © 2019 International Society of Nephrology Terms and Conditions

Figure 3 The effect of β-hydroxybutyrate (β-OHB) on pyroptosis in ischemia-reperfusion (IR) injured kidneys. (a) Representative results of terminal deoxynucleotidyl transferase–mediated dUTP nick end-labeling (TUNEL) staining of the kidney section in each mice group. Arrowheads indicate TUNEL-positive cells. Bars = 50 μm. Semiquantitative analysis of TUNEL-positive cells is shown in right panel. Original magnification ×200. (b–k) The quantitative real-time polymerase chain reaction analysis of mRNA expressions and Western blot analysis of markers related to apoptosis and pyroptosis activation including Bax (b,c), Bcl-2 (d,e), forkhead transcription factor O3 (FOXO3) (f,g), apoptosis repressor with caspase recruitment domain (ARC) (h), caspase-1 (i), as well as the proinflammatory cytokines interleukin-1β (IL-1β) (j) and IL-18 (k) in each mouse group. The experimental groups consist of phosphate-buffered saline–treated mice with sham operation (n = 6), β-OHB–treated sham mice (sham+β-OHB) (n = 6), phosphate-buffered saline–treated mice with IR injury (n = 8), and β-OHB–treated IR mice (IR+β-OHB) (n = 8). ##P < 0.01 versus sham, #P < 0.05 versus sham, **P < 0.01 versus IR, *P < 0.05 versus IR. Bar represents mean ± SEM. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org. Kidney International 2019 95, 1120-1137DOI: (10.1016/j.kint.2018.11.034) Copyright © 2019 International Society of Nephrology Terms and Conditions

Figure 4 The effect of β-hydroxybutyrate (β-OHB) on pyroptosis in hypoxic and H2O2-treated human kidney-2 (HK-2) cells. (a) Cell viability in normoxic and hypoxic HK-2 cells treated with increasing concentrations of β-OHB. ##P < 0.01 versus normoxic HK-2 cells, **P < 0.01 versus hypoxic HK-2 cells without β-OHB, *P < 0.05 versus hypoxic HK-2 cells without β-OHB, ¶¶P < 0.01 versus hypoxic HK-2 cells with 10 mmol/l (mM) of β-OHB, ¶P < 0.05 versus hypoxic HK-2 cells with 10 mM of β-OHB. The quantitative real-time polymerase chain reaction analysis of mRNA expression levels of forkhead transcription factor O3 (FOXO3) (b), apoptosis repressor with caspase recruitment domain (ARC) (c), caspase-1 (d), interleukin-1β (IL-1β) (e), IL-18 (f), Bax (g), and Bcl-2 (h) in hypoxic HK-2 cells. ##P < 0.01 versus normoxic HK-2 cells, **P < 0.01 versus hypoxic HK-2 cells without β-OHB, *P < 0.05 versus hypoxic HK-2 cells without β-OHB. n = 5 in each experimental group. (i) Time-dependent FOXO3 mRNA expression in HK-2 cells treated with 2 mM of H2O2. ##P < 0.01 versus control HK-2 cells, **P < 0.01 versus H2O2-treated HK-2 cells for 3 hours, *P < 0.05 versus H2O2-treated HK-2 cells for 3 hours. n = 4 in each experimental group. (j) The effects of β-OHB (10 mM) on cell viability in H2O2 (2 mM)-treated HK-2 cells. ##P < 0.01 versus control HK-2 cells, **P < 0.01 versus. H2O2 (2 mM)-treated HK-2 cells, *P < 0.05 versus H2O2 (2 mM)-treated HK-2 cells. n = 5 in each experimental group. (k) The effects of β-OHB (10 mM) on mRNA expression levels of FOXO3, ARC, caspase-1, IL-1β, and IL-18 in H2O2-treated HK-2 cells. ##P < 0.01 versus control HK-2 cells, *P < 0.05 versus H2O2 (2 mM)-treated HK-2 cells. n = 5 in each experimental group. Bar represents mean ± SEM. Kidney International 2019 95, 1120-1137DOI: (10.1016/j.kint.2018.11.034) Copyright © 2019 International Society of Nephrology Terms and Conditions

Figure 5 The effect of forkhead transcription factor O3 (FOXO3) knockdown on pyroptosis related genes in human kidney-2 (HK-2) cells. HK-2 cells were transfected with small, interfering RNAs (siRNAs) specific for the human FOXO3 gene or negative control siRNA and subject to hypoxic insults by using hypoxia incubator. The quantitative real-time polymerase chain reaction analysis of mRNA expression levels of FOXO3 (a), apoptosis repressor with caspase recruitment domain (ARC) (b), caspase-1 (c), interleukin-1β (IL-1β) (d), and IL-18 (e). ##P < 0.01 versus normoxic HK-2 cells with control siRNA, #P < 0.05 versus normoxic HK-2 cells with control siRNA, **P < 0.01 versus hypoxic HK-2 cells with control siRNA. *P < 0.05 versus hypoxic HK-2 cells with control siRNA. n = 5 in each experimental group. Bar represents mean ± SEM. Kidney International 2019 95, 1120-1137DOI: (10.1016/j.kint.2018.11.034) Copyright © 2019 International Society of Nephrology Terms and Conditions

Figure 6 The effect of forkhead transcription factor O3 (FOXO3) knockdown on β-hydroxybutyrate (β-OHB)–mediated anti-pyroptosis in hypoxic human kidney-2 (HK-2) cells. The quantitative real-time polymerase chain reaction analysis of mRNA expression levels of apoptosis repressor with caspase recruitment domain (ARC) (a), caspase-1 (b), interleukin-1β (IL-1β) (c), and IL-18 (d), as well as cell viability (e) in β-OHB–treated hypoxic HK-2 cells transfected with small, interfering RNAs (siRNAs) specific for the human FOXO3 gene or negative control siRNA. $$P < 0.01 versus hypoxic HK-2 cells, $P < 0.05 versus hypoxic HK-2 cells, ##P < 0.01 versus β-OHB (10 mmol/l [mM])–treated hypoxic HK-2 cells, **P < 0.01 versus β-OHB (10 mM)–treated hypoxic HK-2 cells transfected with control siRNA. n = 5 in each experimental group. Bar represents mean ± SEM. Kidney International 2019 95, 1120-1137DOI: (10.1016/j.kint.2018.11.034) Copyright © 2019 International Society of Nephrology Terms and Conditions

Figure 7 The effects of β-hydroxybutyrate (β-OHB) on histone acetylation and the activities of histone deacetylases (HDACs)/histone acetyltransferases (HATs) in ischemic-reperfusion (IR)–injured kidneys and hypoxic human kidney-2 (HK-2) cells. (a) The expression of acetyl-histone H3 and histone H3 in each mouse group; phosphate-buffered saline–treated mice with sham operation (n = 6), β-OHB–treated sham (sham+β-OHB) (n = 6), phosphate-buffered saline–treated mice with IR injury (n = 8), and β-OHB–treated IR (IR+β-OHB) (n = 8). The lower panel represents the quantification of band intensity. (b) HDAC activities (left panel) and HAT activities (right panel) in each mouse group. (a,b) ##P < 0.01 versus sham, #P < 0.05 versus sham, *P < 0.05 versus IR. (c) The expression of acetyl-histone H3 in hypoxic HK-2 cells with or without increasing concentrations of β-OHB. The lower graph represents the ratio of acetyl histone H3 to histone H3. (d) The activities of HDACs (left panel) and HATs (right panel) in hypoxic HK-2 cells treated with β-OHB. (c,d) ##P < 0.01 versus normoxic HK-2 cells, **P < 0.01 versus hypoxic HK-2 cells without β-OHB, *P < 0.05 versus hypoxic HK-2 cells without β-OHB. n = 5 in each group. (e) Chromatin immunoprecipitation assay for acetylated H3K9 levels in the forkhead transcription factor O3 (FOXO3) promoter. The percentage of input DNA values were enrichment levels of immunoprecipitated promoter fragments determined by quantitative polymerase chain reaction normalized to the level of input (nonimmunoprecipitated) DNA. #P < 0.05 versus normoxic HK-2 cells, **P < 0.01 versus hypoxic HK-2 cells without β-OHB. n = 4 in each group. Bar represents mean ± SEM. AcH3K9, lysine-9 acetylation in histone H3; mM, mmol/l. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org Kidney International 2019 95, 1120-1137DOI: (10.1016/j.kint.2018.11.034) Copyright © 2019 International Society of Nephrology Terms and Conditions

Figure 8 The effects of β-hydroxybutyrate (β-OHB) on pyroptosis in ischemic-reperfusion (IR)–injured kidneys in the case that continuous infusion is stopped just after renal ischemic insult. (a) β-OHB plasma levels were measured 24 hours before and after the sham or IR operation in each mouse group; phosphate-buffered saline–treated sham-operated mice, phosphate-buffered saline–treated IR-operated mice, IR where β-OHB treatment was stopped just after IR operation (IR+β-OHB [24 h]), and IR treated with β-OHB all through the experimental period (IR+β-OHB [48 h]). Blood urea nitrogen levels (b) and serum levels of creatinine (c) were measured 24 hours after the sham or IR operation. The quantitative real-time polymerase chain reaction analysis of mRNA expressions of markers related to pyroptosis activation including forkhead transcription factor O3 (FOXO3) (d), apoptosis repressor with caspase recruitment domain (ARC) (e), caspase-1 (f), as well as the proinflammatory cytokines interleukin-1β (IL-1β) (g) and IL-18 (h). (i) Western blot analysis of acetyl-histone H3 and histone H3 in each mice group. The lower panel represents the quantification of band intensity. ##P < 0.01 versus sham, **P < 0.01 versus IR, *P < 0.05 versus IR, $$P < 0.01 versus IR+β-OHB (24 h), and $P < 0.05 versus IR+β-OHB (24 h). n = 4 for each mouse group. Bar represents mean ± SEM. AcH3K9, lysine-9 acetylation in histone H3; GAPDH, glyceraldehyde-3-phosphate dehydrogenase. Kidney International 2019 95, 1120-1137DOI: (10.1016/j.kint.2018.11.034) Copyright © 2019 International Society of Nephrology Terms and Conditions

Figure 9 Schema depicting the mechanisms for renoprotection by β-hydroxybutyrate (β-OHB) against ischemia-reperfusion (IR) injury. In IR injuries, both histone acetyltransferase (HAT) and histone deacetylase (HDAC) activities decreased although the decrease in HAT activity was more responsible for decreased levels of acetylated H3K9. These changes might downregulate forkhead transcription factor O3 (FOXO3) and apoptosis repressor with caspase recruitment domain (ARC) leading to upregulate caspase-1, interleukin-1β (IL-1β), and IL-18. These gene alterations resulted in pyroptosis and renal tubular injuries after IR insults. The treatment with β-OHB inhibited HDAC activity and further reduced the HDAC activity in the IR-injured kidney and ameliorated the decreased levels of acetylated H3K9. This effect by β-OHB reversed the pyroptosis gene alterations in the IR-injured kidney and ameliorated renal dysfunction. Kidney International 2019 95, 1120-1137DOI: (10.1016/j.kint.2018.11.034) Copyright © 2019 International Society of Nephrology Terms and Conditions

Figure S1 The effects of β-hydroxybutyrate (β-OHB) on pyroptosis in IR-injured kidneys at different time point. (A–E) The quantitative real-time polymerase chain reaction analysis of mRNA expressions of markers related to pyroptosis activation including forkhead transcription factor O3 (FOXO3) (A), apoptosis repressor with caspase recruitment domain (ARC) (B), caspase-1 (C), as well as the proinflammatory cytokines interleukin-1β (IL-1β) (D) and IL-18 (E) in mice groups 8 hours or 24 hours after reperfusion. The experimental groups consist of phosphate-buffered saline–treated mice with sham operation (sham, n = 5), phosphate-buffered saline–treated mice with ischemia reperfusion (IR) injury (n = 5), and IR treated with β-OHB (IR+β-OHB) (n = 5) for 8 hours or 24 hours after reperfusion. ##P < 0.01 versus sham, #P < 0.05 versus sham, **P < 0.01 versus IR, *P < 0.05 versus IR, $$P < 0.01 versus IR (8 hours after reperfusion), $P < 0.05 versus IR (8 hours after reperfusion), ¶¶P < 0.01 versus IR+β-OHB (8 hours after reperfusion). Bar represents mean ± SEM. Kidney International 2019 95, 1120-1137DOI: (10.1016/j.kint.2018.11.034) Copyright © 2019 International Society of Nephrology Terms and Conditions