Hepatocyte nuclear factor 4α mediated quinolinate phosphoribosylltransferase (QPRT) expression in the kidney facilitates resilience against acute kidney injury

Abstract

Nicotinamide adenine dinucleotide (NAD+) levels decline in experimental models of acute kidney injury (AKI). Attenuated enzymatic conversion of tryptophan to NAD+ in tubular epithelium may contribute to adverse cellular and physiological outcomes. Mechanisms underlying defense of tryptophan-dependent NAD+ production are incompletely understood. Here we show that regulation of a bottleneck enzyme in this pathway, quinolinate phosphoribosyltransferase (QPRT) may contribute to kidney resilience. Expression of QPRT declined in two unrelated models of AKI. Haploinsufficient mice developed worse outcomes compared to littermate controls whereas novel, conditional gain-of-function mice were protected from injury. Applying these findings, we then identified hepatocyte nuclear factor 4 alpha (HNF4α) as a candidate transcription factor regulating QPRT expression downstream of the mitochondrial biogenesis regulator and NAD+ biosynthesis inducer PPARgamma coactivator-1-alpha (PGC1α). This was verified by chromatin immunoprecipitation. A PGC1α - HNF4α -QPRT axis controlled NAD+ levels across cellular compartments and modulated cellular ATP. These results propose that tryptophan-dependent NAD+ biosynthesis via QPRT and induced by HNF4α may be a critical determinant of kidney resilience to noxious stressors.

Keywords: HNF4α; NAD; acute kidney injury; metabolism.

Figure 1:. De novo NAD+ Biosynthesis suppression is a component of nephrotoxic AKI.

a. Schema of NAD+ Biosynthesis. b. Volcano plot of mouse kidney metabolites from mice that received cisplatin or vehicle. P-values calculated using Mann-Whitney. Dotted lines indicate p value greater of 0.05 (horizontal) and fold change < or > 2 (vertical). c. Heat map showing individual values of selected metabolites. Values are normalized to average of vehicle animals and presented as fold change. Darkest colors indicate fold change ≥6 (red) or ≤-6 (blue). d. Heatmap representing gene expression of selected genes in WT male mice who received 25mg/kg IP cisplatin (n=17) or Vehicle (n=10). Data are normalized to the vehicle average and expressed as fold change. Darkest colors indicate fold change ≥5 (red) or ≤-5 (blue). e. Correlation of kidney QPRT expression to kidney Lcn2 expression in WT mice that received cisplatin or vehicle. Correlation and p-value calculation using a simple linear regression, f. Correlation of 3 kidney injury markers to kidney QPRT mRNA expression. Correlation and p-values calculated with simple linear regression.

Figure 2:. QPRT expression mediates NAD+ and ATP.

a. Schematic describing a compartment-specific NAD+ biosensor. b-d Representative images of compartment-specific NAD+ biosensor in use. Relative cytoplasmic (e), nuclear (f), and mitochondrial (g) NAD+ in QPRT overexpression (OE), control, and siQPRT knock down in HK-2 cells. h. Relative ATP in QPRT OE and Control I. Relative ATP in siQPRT and Control. P-values were calculated using Mann-Whitney. * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

Figure 3:. QPRT +/− mice are more susceptible to two distinct models of nephrotoxic AKI.

a. Schema of Cisplatin AKI experiment. b. Serum creatinine, c. serum BUN, and d. renal mRNA expression of Lcn2 after cisplatin (72 hours after 25mg/kg IP) in QPRT +/+ (n=23) and QPRT +/− (n=25) littermates. All compared with Mann-Whitney. e. Tubular injury scoring from histology. f. Representative PAS images after cisplatin. g. Schema of folic acid AKI experiment. h. Serum creatinine, i. renal mRNA expression of Lcn2, j. serum BUN, and k. tubular injury scoring after folic acid (24 hours after 250mg/kg IP) in QPRT +/+ (n=24) and QPRT +/− (n=21) littermates. l. Representative PAS images after folic acid. Dotted arrows represent dilated tubules with intratubular casts. Solid arrows represent necrotic cells and denuded tubules. Black scale bars represents 100um. * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

Figure 4:. iNephQPRT mice are protected against two distinct models of nephrotoxic AKI.

a. Schema of iNephQPRT mouse generation. b. Schema of iNephQPRT cisplatin experiment. c. Serum BUN, renal lcn2 mRNA (d), and serum creatinine (e) in iNephQPRT (n=19) and control (n=13) littermates after cisplatin (72 hours after cisplatin 25mg/kg IP). f. Correlation of serum BUN with QPRT mRNA expression. g. Schema of folic acid experiment. Serum BUN (h), renal lcn2 mRNA (i), serum creatinine (j), and tubular injury scoring after folic acid (24 hours after 250mg/kg IP) (k) in iNeph QPRT mice (n=23) and control littermates (n=27). l. Representative histology after folic acid iNephQPRT mice and controls. Dotted arrows represent dilated tubules with intratubular casts. Solid arrows represent necrotic cells and denuded tubules. Black scale bars represent 100um.* = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

Figure 5:. Bioinformatic analysis proposed HNF4α as a transcription factor linking PGC1α and QPRT.

a. Schematic representing role of unknown transcription factor linking PGC1α and QPRT. b. Venn diagram comparing transcription factors known to bind QPRT with transcription factors activated by PGC1α.

Figure 6:. HNF4α activity mirrors QPRT activity.

a. In human kidney snRNASeq, PGC1α is ubiquitously expressed, but HNF4α localizes to the proximal tubule. b. De novo NAD+ biosynthesis and QPRT also localize to the proximal tubule (left of dotted line), while enzymes of other NAD+ biosynthetic pathways are expressed throughout the kidney. c. After cisplatin, the degree of QPRT suppression correlates with HNF4α suppression in mouse kidneys. d. Overexpressing HNF4α (OE) increases QPRT expression. HNF4α OE increases NAD+, while siHNF4a decreases NAD+ in the cytoplasmic (e) nuclear (f), and mitochondrial (g) compartments. h. HNF4α OE increases ATP, and siHNF4α decreases ATP (I). * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

Figure 7:. HNF4α modulates renal QPRT, connecting PGC1α to de novo NAD+ biosynthesis.

a. QPRT qPCR of HNF4α ChIP confirms that HNF4α does enrich at the QPRT locus in kidney tissue. b. In iNephPGC1α mice, PGC1α overexpression increases HNF4α enrichment at QPRT, and (c) the level of enrichment correlates with the degree of PGC1α overexpression. d. siHNF4α abrogates the PGC1α mediated increase in QPRT expression. * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.