Organ Protection by Caloric Restriction Depends on Activation of the De Novo NAD+ Synthesis Pathway

Abstract

Significance statement: AKI is a major clinical complication leading to high mortality, but intensive research over the past decades has not led to targeted preventive or therapeutic measures. In rodent models, caloric restriction (CR) and transient hypoxia significantly prevent AKI and a recent comparative transcriptome analysis of murine kidneys identified kynureninase (KYNU) as a shared downstream target. The present work shows that KYNU strongly contributes to CR-mediated protection as a key player in the de novo nicotinamide adenine dinucleotide biosynthesis pathway. Importantly, the link between CR and NAD+ biosynthesis could be recapitulated in a human cohort.

Background: Clinical practice lacks strategies to treat AKI. Interestingly, preconditioning by hypoxia and caloric restriction (CR) is highly protective in rodent AKI models. However, the underlying molecular mechanisms of this process are unknown.

Methods: Kynureninase (KYNU) knockout mice were generated by Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) and comparative transcriptome, proteome and metabolite analyses of murine kidneys pre- and post-ischemia-reperfusion injury in the context of CR or ad libitum diet were performed. In addition, acetyl-lysin enrichment and mass spectrometry were used to assess protein acetylation.

Results: We identified KYNU as a downstream target of CR and show that KYNU strongly contributes to the protective effect of CR. The KYNU-dependent de novo nicotinamide adenine dinucleotide (NAD+) biosynthesis pathway is necessary for CR-associated maintenance of NAD+ levels. This finding is associated with reduced protein acetylation in CR-treated animals, specifically affecting enzymes in energy metabolism. Importantly, the effect of CR on de novo NAD+ biosynthesis pathway metabolites can be recapitulated in humans.

Conclusions: CR induces the de novo NAD+ synthesis pathway in the context of IRI and is essential for its full nephroprotective potential. Differential protein acetylation may be the molecular mechanism underlying the relationship of NAD+, CR, and nephroprotection.

Graphical abstract

Figure 1

KYNU is expressed predominantly in the proximal tubule and induced by two modes of preconditioning. (A) Schematic illustration of the experimental setup modified from Johnsen et al. JASN 2020. C57Bl/6J mice were divided into three treatment groups (non-PC, HP, and CR) before IRI. Right nephrectomy was performed at the time point of surgery (0 hour). To obtain blood and the left kidneys, 24 hours after reperfusion, all mice were sacrificed. (B) Venn diagram illustrating genes only regulated by CR (3359), HP (91), or both (230) at 0 hour before IRI modified from Johnsen et al. JASN 2020 (shiny.cecad.uni-koeln.de:3838/IRaP/). Preconditioning-induced expression changes of Kynu pre-IRI (CR versus non-PC and HP versus non-PC) are displayed in the box. (C) Schematic illustration of the segmental localization of Kynu along the nephron compiled from previously published single-cell, single-nucleus, and bulk RNAseq data as well as proteome analyses of microdissected kidneys. This illustration was created with BioRender.com (D) Immunohistochemical staining of KYNU in murine kidneys confirms the predominant localization in the proximal tubule. (E) Immunohistochemical staining of an adult human kidney using anti-KYNU antibody (HPA031686, derived from www.humanproteinatlas.org ). Ascending thin limb of Henle loop; connecting tubule; cortical collecting duct; cortical thick ascending limb; d, day; descending thin limb of Henle loop, long-loop, inner medulla; descending thin limb of Henle loop, short-loop; descending thin limb of Henle loop, long-loop, outer medulla; distal convoluted tubule; HP, hypoxic preconditioning; inner medullary collecting duct; medullary thick ascending limb; outer medullary collecting duct; padj: adjusted P value; S1 Proximal tubule; S2 Proximal tubule; S3 Proximal tubule.

Figure 2

Generation and basal phenotyping of conventional KYNU^null mice. (A) Schematic illustration of the strategy to generate KYNU^null mice by Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)/Cas9-mediated NHEJ. (B) Immunoblot of murine kidneys showing the successful deficiency in KYNU^null mice compared with expression in WT mice (loading control: Pan-14-3-3). (C) Histologic analyses by staining with anti-KYNU in kidneys of KYNU^null mice confirming the deficiency compared with kidneys of WT (Figure 1D). (D) Bodyweight of KYNU^null mice and WT littermates recorded for 1 year at the age of 4, 8, 12, 16, 26, and 52 weeks (n=5 per group). (E) Plasma creatinine values of KYNU^WT and KYNU^null mice; t test, n=5 per group. (F) Plasma values of BUN of KYNU^WT mice and KYNU^null mice; t test, n=5 per group. (G) PAS staining of kidneys of a 14-week-old WT mouse and a 14-week-old KYNU^null mouse.n. s.: P value≥0.05. Bars indicate means±SD; blue, CRISPR-RNA sequence; bold capitals, amino acids; green, first methionine; KYNU^null, KYNU deficiency; KYNU^WT: WT; red, protospacer adjacent motif side.

Figure 3

KYNU deficiency does not increase kidney damage after IRI but diminishes the protective potential of preconditioning. (A) Plasma creatinine values 24 hours after IRI. IRI induces a similar increase of plasma creatinine in KYNU^null and WT mice. KYNU deficiency leads to a reduction of the protection by CR. Non-PC and CR were analyzed using one-way ANOVA and Tukey post hoc test for multiple comparisons. For the comparison of CR-KYNU^WT and CR-KYNU^null mice, the t test was used. (B) Plasma values of BUN24 hours after IRI recapitulate the findings for plasma creatinine. Non-PC and CR were analyzed using one-way ANOVA and Tukey post hoc test for multiple comparisons. For the comparison of CR-KYNU^WT and CR-KYNU^null mice, the t test was used. (C) PAS of kidneys 24 hours after IRI. #: protein casts; square: pyknosis; black arrowhead: loss of brush border; gray arrowhead: epithelial flattening. (D) Analysis of cell death 24 hours after IRI using TUNEL. CR strongly protect from cell death while this potential is partly lost in KYNU^null mice. (E) Semiquantitative composite damage score; one-way ANOVA and Tukey post hoc test (n=10–12 per group). (F) Quantification of TUNEL-positive area to total area. The t test was used (n=3 per group). ****P value <0.0001; **P=0.001–0.01; *P value <0.05; n.s.: P value ≥0.05. Bars indicate means±SD; KYNU^null, KYNU deficiency; KYNU^WT, WT; scale bars indicate 100 µm; sham: right nephrectomy followed by no clamping of the left renal pedicle.

Figure 4

CR modulates the Trp metabolic pathway in a KYNU-dependent manner. (A) Illustration of the Trp metabolic pathway as one of the three NAD+ biosynthesis pathways. (1) de novo; (2) Preiss-Handler; (3) belonging to 1, 2, and 4; and (4) salvage. (B) Quantification of metabolites in the Trp metabolic pathway comparing undamaged kidneys of KYNU^WT and KYNU^null mice (unpaired t test, n=5 per group). Only metabolites that show a significant change are shown in this graph (for others, see Supplemental Figure S5). (C) Quantification of metabolites involved in the Trp metabolic pathway comparing WT kidneys from the non-PC and CR groups (unpaired t test, n=5 per group). Only metabolites that show a significant change are given (for others, see Supplememtal Figure S5). (D) Quantification of metabolites involved in the Trp metabolic pathway comparing analyses of blood plasma from the non-PC and CR groups (unpaired t test, n=10 per group). Only metabolites that show a significant change are given (for others, see Supplemental Figure S5). (E) Quantification of metabolites involved in the Trp metabolic pathway comparing longitudinally acquired human serum samples pre-CR and post-CR (paired t test, n=15 per group). Only metabolites that show a significant change are given (for others, see Supplemental Figure S5). ***P value <0.001; **P value: 0.001–0.01; *P value <0.05; n.s.: P value ≥0.05. (1) de novo branch of NAD+ biosynthesis; (2) Preiss-Handler branch of NAD+ biosynthesis; (3) belonging to 1, 2, and 4; and (4) salvage branch of NAD+ biosynthesis; AA, anthranilic acid; ACMS, α-amino-beta-carboxymuconate-epsilon-semialdehyde; bars indicate means±SD; KYNU^WT, WT; KYNU^null, KYNU deficiency; NAAD, nicotinic acid adenine dinucleotide; NADSYN1, NAD synthetase 1; N-formyl-kyn, N-formyl-Kynurenine; NAM, nicotinamide; NAMN, nicotinic acid mononucleotide; NAMPT, nicotinamide phosphoribosyltransferase; NMNAT 1/2/3, nicotinamide mononucleotide adenylyltransferase 1/2/3; NMN, nicotinamide mononucleotide.

Figure 5

CR induces profound changes in transcript and protein levels of genes involved in NAD+ biosynthesis and maintains NAD+ levels after renal IRI. (A) Heatmap depicting mean-based z-score transcript levels (RNAseq) of the genes involved in the NAD+ biosynthesis pre-IRI and at 24 hours post-IRI in non-PC mice or after CR; asterisks reflect significance of pairwise comparisons of each gene between treatment groups (n=5 per group). (B) Heatmap depicting median-based z-score protein expression levels (LC-MS/MS) of the enzymes involved in NAD+ biosynthesis (selection because not all proteins were detected by LC-MS/MS) at 0 hour before (pre-IRI) and at 4 hour after IRI (post-IRI) in non-PC mice or after CR; asterisks reflect significance of pairwise comparisons of each gene between treatment groups (n=5 per group). (C) Quantification of 3OH-L-kyn comparing damaged (post-IRI) kidneys of KYNU^wt and KYNU^null mice without preconditioning (non-PC) or after CR. (D) NAD levels of right, undamaged and left, damaged kidneys of KYNU^WT and KYNU^null mice without previous preconditioning (non-PC) or after CR (t test, n=6 per group). The direct comparison of CR-KYNU^wt and CR-KYNU^null does not reach significance (t test: P value: 0.2; n=6 per group). n. s.: P value >0.05, *adjusted P value/q value <0.05–0.01, ***adjusted P value/q value <0.0001. (1) de novo branch of NAD+ biosynthesis; (2) Preiss-Handler branch of NAD+ biosynthesis; (3) belonging to 1, 2 and 4; and (4) salvage branch of NAD+ biosynthesis; NAD: total NAD (NAD+ and NADH).

Figure 6

CR prevents acetylation of metabolic key targets after IRI. (A) Principal component analysis of the acetyl-lysine sites. (B) Volcano plot depicting acetyl-lysine sites of not preconditioned kidney post-IRI in comparison with not preconditioned kidneys pre-IRI (paired two-sided t test, S₀=0.1, permutation-based false discovery rate (FDR)=0.1, 500 randomizations, n=5 per group). (C) Volcano plot depicting acetyl-lysine sites of CR kidneys post-IRI in comparison with CR kidneys pre-IRI (paired two-sided t test, S₀=0.1, permutation-based FDR=0.1, 500 randomizations, n=5 per group). (D) Volcano plot depicting acetyl-lysine sites of CR kidneys post-IRI in comparison with non-PC kidneys post-IRI (unpaired two-sided t test, S₀=0.1, permutation-based FDR=0.1, 500 randomizations, n=5 per group). (E) Kyoto Encyclopedia of Genes and Genomes terms of differentially downregulated acetyl-lysine sites from the comparison of CR kidneys post-IRI in comparison with non-PC kidneys post-IRI. (F) Schematic illustration of the metabolic pathways affected by less acetylation in the comparison of CR kidneys post-IRI in comparison with non-PC kidneys post-IRI. ARA, arachidonic acid; DHETs, dihydroxyeicosatrienoic acids; EETs, epoxyeicosatrienoic acids; post-IRI, 24 hours after IRI; pre-IRI: right kidneys withdrawn directly before IRI.