The Significance of NAD+ Biosynthesis Alterations in Acute Kidney Injury

Summary

Acute kidney injury (AKI) is a serious and highly prevalent disease, yet only supportive treatment is available. Nicotinamide adenine dinucleotide (NAD+) is a cofactor necessary for adenosine triphosphate (ATP) production and cell survival. Changes in renal NAD+ biosynthesis and energy utilization are features of AKI. Targeting NAD+ as an AKI therapy shows promising potential. However, the pursuit of NAD+-based treatments requires deeper understanding of the unique drivers and effects of the NAD+ biosynthesis derangements that arise in AKI. This article summarizes the NAD+ biosynthesis alterations in the kidney in AKI, chronic disease, and aging. To enhance this understanding, we explore instances of NAD+ biosynthesis alterations outside the kidney in inflammation, pregnancy, and cancer. In doing so, we seek to highlight that the different NAD+ biosynthesis pathways are not intercon-vertible and propose that the way in which NAD+ is synthesized may be just as important as the NAD+ produced.

Acute kidney injury (AKI) affects one in five hospitalized adults and one in three hospitalized children worldwide, yet there is still no treatment.¹ An extensive body of literature implicates aberrations in renal tubular metabolism as critical drivers of AKI, yet the search for means to restore normal metabolism has been a challenge. Among these metabolic derangements, nicotinamide adenine dinucleotide (NAD+) biosynthesis is dysregulated in AKI. This contributes to decreased renal NAD+ and accumulation of specific NAD+ precursors. Restoring renal NAD+ through alternate biosynthetic pathways protects against AKI.² Yet, the physiologic drivers of this seemingly maladaptive attenuation of normal biosynthetic flux remain to be elucidated.

NAD+ is synthesized by three pathways, as follows: from dietary niacin (NA) via the Preiss-Handler (PH) pathway, dietary tryptophan (TRP) via the de novo biosynthesis pathway, and dietary or recycled nicotinamide (NAM) via the salvage pathway (Fig. 1). The liver is the body’s main source of NAD+ precursors. Dietary NA, TRP, and NAM are absorbed from the gut and taken to the liver for metabolism. There, nearly all precursors are absorbed from circulation and converted to NAD+. To maintain tight control of NAD+ level, most NAD+ is subsequently converted to NAM. After meeting hepatic NAD+ needs, excess NAM is returned to the circulation for systemic use. Most tissues in the body use circulating NAM to meet tissue-specific NAD + needs via the salvage pathway alone. Few nonhepatic tissues can make additional small amounts of NAD+ from NA. The kidney is the only other organ to express all three NAD+ biosynthetic pathways and the only nonhepatic organ to generate excess NAM for systemic export (Fig. 2).³ However, the kidney contributes roughly 5% of total circulating NAM, so renal NAD+ production is unlikely to be a critical driver of systemic NAD+ needs.

Figure 1.

Pathways of nicotinamide adenine dinucleotide (NAD+) biosynthesis. Dotted lines represent theoretical pathways. Abbreviations: AFMID, arylformamidase; IDO, indoleamine 2,3-dioxygenase; KMO, kynurenine 3-monooxygenase; KYNU, kynureninase; NAAD, nicotinic acid adenine dinucleotide; NADSYN, nicotinamide adenine dinucleotide synthetase; NAMN, nicotinic acid mononucleotide; NAMPT, nicotinamide phosphoribostransferase; NAPRT, nicotinate phosphoribosyltransferase; NMN, nicotinamide mononucleotide; NMNAT, nicotinamide nucleotide adenylyltansferase; NRK1,2, nicotinamide riboside kinase 1,2; PNP, purine-nucleoside phosphorylase; QPRT, quinolinic acid phosphoribosyltransferase; TDO, tryptophan-2,3-dioxygenase; 3HAO, 3, hydroxyanthranilate 3,4 dioxygenase.

Figure 2.

Different organs use different sources to synthesize nicotinamide adenine dinucleotide (NAD+). Abbreviations: NA, nicotinic acid; NAM, nicotinamide.

Therefore, one must ask why the kidney has evolved to use all NAD+ biosynthetic pathways and synthesize NAD + in excess? Understanding this phenomenon may offer crucial insight into the kidney’s metabolic needs, the effect of acute stress on those needs, and the role of NAD+ biosynthetic pathways beyond just increasing NAD+ levels. This review attempts to broaden the understanding of NAD + biosynthesis and stress-driven biosynthetic change in the kidney by examining renal and nonrenal changes in NAD+ biosynthesis in the literature. By exploring well-described instances of NAD+ biosynthesis change in inflammation, pregnancy, and cancer, we hope to introduce new ways to think about NAD+ metabolism in AKI.

NAD+ BIOSYNTHESIS ALTERATIONS IN THE KIDNEY

AKI

Renal NAD+ levels decrease in ischemic, septic, and chemotoxic AKI. Restoring NAD+ levels via NAD+ precursor supplementation may protect against AKI.^2,4–6 NAD+ consumption by injury-responsive enzymes and transcription factors such as poly (adenosine diphosphate-ribose) polymerases (PARPs) and sirtuins is altered in AKI. However, the physiological consequences of these activities may not consistently relate to cellular NAD+ availability. For example, inhibiting PARP1 increased NAD+ and protected against AKI.⁷ However, inhibition of SIRT3 activity also increased NAD+ but exacerbated AKI.^8,9 Similarly, SIRT1 overexpression or pharmacologic activation protected against AKI but decreased total NAD+.^10,11 Given opposing physiological consequences of renal NAD+ consumers, it is likely that increased NAD+ consumption is not solely responsible for the pathologic NAD+ derangements seen in AKI.

Indeed, biosynthesis of NAD+ also is altered in AKI.^2,4,6,12,13 Although suppression of all three pathways of NAD+ biosynthesis has been shown, the de novo biosynthesis pathway shows the most profound suppression and metabolite accumulation.^2,14 De novo biosynthetic pathway metabolites accumulate to such a degree in AKI that they frequently have appeared in nonbiased metabolomic screening of urine, blood, or tissue from experimental and human AKI.^2,13 A recent study even showed that the urinary quinolinic acid (QA)/TRP ratio measured in human beings one day after cardiac bypass surgery was a better predictor of postoperative AKI than serum creatinine or blood urea nitrogen levels.¹³ Another study reported that up-regulation of the de novo NAD+ biosynthesis pathway enzymes was one of the most abundant transcriptomic and proteomic changes noted in kidneys preconditioned with calorie restriction or hypoxia to better withstand acute stress.¹⁴ These preconditioned animals had normal NAD+ levels when uninjured, but, upon injury, they showed less NAD+ reduction and less severe injury. Variation in NAD+ biosynthesis, particularly the de novo biosynthesis pathway, therefore may be a factor contributing to variation in renal stress resilience.

CHRONIC KIDNEY DISEASE AND FIBROSIS

Because NAD+ is critical for metabolic processes in every cell, it frequently has been studied in aging-related diseases.^15–18 Aging in the normal kidney shares structural and functional similarities with chronic kidney disease (CKD), including increased fibrosis and scarring, reduced microvessel density, and decreased glomerular filtration rate.¹⁹ Recent studies have associated down-regulation of de novo NAD+ enzymes and decreased renal and serum NAD+ levels with CKD progression.^18,20,21 The expression of nearly every enzyme in this pathway is suppressed, including the bottleneck enzyme, quinolinate phosphoribosyltransferase (QPRT).²¹ Persistent reduction of QPRT expression after AKI has been linked to CKD progression.¹³ There also may be salvage pathway suppression, but such findings are variable.^6,21

In contrast to AKI, in which the therapeutic potential of NAD+ has been shown by several independent groups, fewer studies have focused on its role in CKD. In an adenine-induced CKD model²² and a unilateral urinary tract obstruction (UUO) model,²³ prophylactic NAM therapy reduced inflammation and oxidative stress and prevented CKD progression. However, administration of NAM to mice with advanced CKD after adenine exposure did not provide benefit.²² In two other CKD experiments examining UUO and progressive tubulointerstitial fibrosis, prophylactic nicotinamide riboside (NR) supplementation did not prevent CKD progression.²¹

Thus, although it does seem as though NAD+ loss may play a role in renal fibrosis and CKD development, it is much less understood than the role of NAD+ in AKI. This may be because protective mechanisms in kidneys may change with chronicity as tissue adaptation or cell death takes place. Early intervention is key to slow disease progression. Preventive strategies based on restoring renal NAD+ at early time points should be investigated.

RENAL AGING

In addition to CKD, normal renal aging is associated with reduced cellular NAD+ levels.²⁴ The cellular actors implicated in this phenomenon are not well characterized. One possible mechanism is accumulation of senescent cells, which secrete bioactive factors, proinflammatory cytokines, and proteases to induce a chronic low-grade inflammation termed inflammaging.²⁵ Inflammaging increased the expression of NAD+ consumers such as CD38, PARP, and sirtuins.^24,26 NAD+ biosynthetic pathways also were altered in aged kidneys.²⁷ The expression of de novo NAD+ biosynthesis pathway enzymes tryptophan 2,3-dioxygenase (TDO), indoleamine 2,3-dioxygenase (IDO), and kynurenine 3-monooxygenase were decreased in aged rat kidneys, while metabolites such as TRP and kynurenine were increased.^27,28 Salvage pathway enzymes nicotinamide phosphoribotransferase (NAMPT) and nicotinamide nucleotide adenyltransferase also were suppressed in aged mice compared with young mice.²⁹

Targeting NAD+ metabolism may be a therapeutic approach to mitigate or reverse age-related disease in kidneys.^29–31 Nicotinamide mononucleotide (NMN) application to human proximal tubule cells decreased expression of senescence-associated β-galactosidase.³² Supplementing aged mice with NMN protected them against age-related cisplatin AKI susceptibility.²⁹ In addition, NMN supplementation restored youthful proteostasis in aged kidneys.³¹ Kidney NAD+ decreases with aging via enhanced consumption and altered production. Although few studies have studied NAD+ targeting as an anti-aging approach in the kidney, supplementing with NAD+ precursors may be an effective strategy for preventing aging-associated renal changes.

REGULATION OF NAD+ BIOSYNTHESIS

Extreme variation exists in NAD+ biosynthesis patterns among cell types and tissues, which complicates a comprehensive understanding of NAD+ regulation. Because the liver most robustly uses all three NAD+ biosynthetic pathways, its NAD+ regulation has been most studied. Early rodent studies showed that the liver tightly regulates NAD+ levels via hydrolysis of NAD+ to NAM rather than NAD+ biosynthesis restriction.³³ Indeed, the Michaelis constant (Km) of QPRT is nearly 100 times higher than the QA concentration in the liver, indicating avid and complete conversion toward NAD+, even with dietary surplus. However, the maximum velocity (Vmax) of QPRT is relatively low, so although all QA should be shuttled to NAD+ under physiologic conditions, transient accumulation of QA is predicted when TRP or other upstream metabolites increase. NAD+ biosynthesis from NA is more regulated. Hepatic nicotinate phosphoribosyltransferase (NAPRT), which converts NA to NAD+, functions at maximum capacity. Therefore, surplus NA does not increase NAD+ as TRP does.³³ However, in vitro studies in liver and kidney cells and enzyme kinetic experiments using recombinant proteins showed that NAD+ production from NA via NAPRT was not inhibited by excess NAD+, so NA catabolism conceivably would continue at a constant rate regardless of NAD+ concentration. In contrast, salvage pathway NAD+ production via NAMPT shows enzyme inhibition at even physiologic NAD+ levels, indicating that the liver would only synthesize NAD+ from NAM if both TRP and NA were deficient.³⁴ Tracing studies have validated this by indicating that the liver would only synthesize NAD+ from NAM if both TRP and NA were deficient. The opposite was seen in simultaneously tested ovaries, lung, heart, and brain.³⁵ NAM, therefore, acts as a storage pool of NAD+ precursor in the liver and for systemic export. Excess NAM leads to increased methyl-NAM, the methylated byproduct produced by nicotinamide N-methyl transferase (NNMT), which is water soluble and excreted in the urine.³³

Studies in isolated perfused kidney and using isotope tracing of NAD+ precursors showed similar findings as hepatic experiments.^3,36 Namely, the kidney imported TRP, NA, and NAM from circulation and used all three precursors to synthesize NAD+. Similar to the liver, the kidney much more readily synthesized NAD+ from NA and TRP and exported excess NAM into circulation. When injected with labeled NA, the kidney rapidly synthesized significant NAD+, in contrast to most other organs.³⁵ Interestingly, perfusion of excess NAM led to excretion of only NAM—no methyl-NAM—in contrast to the liver. This may indicate that NNMT is less active in the kidney, at least in the uninjured state, although renal NNMT activity is of rapidly increasing interest.³⁷

Given that the liver and kidney prefer using TRP and NA as NAD+ precursors and may have poor salvage pathway function, it has been questioned how salvage pathway NAD+ precursors may have a clinical effect. The best hypotheses for this conundrum have been proposed by Trammel et al³⁸ in a series of metabolomic and tracing studies examining NR metabolism. They showed that NR supplementation increases NAD+ via nicotinic acid adenine dinucleotide, indicating a bypass of the salvage pathway. Labeled NR led to labeled nicotinic acid adenine dinucleotide, followed by labeled NAD+ and NAM. This indicated that there may have been accessory reactions in the setting of high NAD+ metabolism or precursor availability that led to deamidation of NMN to nicotinic acid mononucleotide (NAMN), which then was synthesized to NAD+ using nonsalvage pathway enzymes.

In summary, the liver avidly converts TRP and NA to NAD+ and acts as the body’s main producer of NAM, which most tissues use to create NAD+. The kidney has evolved the same patterns of NAD+ biosynthesis despite not significantly contributing to systemic NAM needs. Under physiologic conditions, these pathways will proceed independently of tissue NAD+ level. There also may be alternate pathways in which salvage pathway precursors can increase NAD+ levels via alternate enzymes. If NAD+ is in surplus, biosynthesis from TRP and NA will not be inhibited. Rather, excess NAD+ will be hydrolyzed to NAM, where it can remain in storage form or be methylated by NNMT for excretion. The hydrolysis of NAD+ also carries clinical implications, as reviewed elsewhere.³⁹ The enzymes that hydrolyze NAD+, namely PARPs, sirtuins, and CD38, are all regulated by various systemic cues to perform broad transcriptional regulation and protein modifications. Understanding why the kidney has evolved tripartite NAD+ biosynthetic capabilities may be critical to elucidating both the seemingly maladaptive role these pathways play in acute stress.

NAD+ BIOSYNTHESIS ALTERATIONS IN NONRENAL PHYSIOLOGY AND PATHOLOGY

Although NAD+ biosynthesis, including de novo NAD+ biosynthesis, is critical to normal renal function, and profound biosynthesis suppression occurs in acute stress, the driving factors are not known. To consider possible benefits or evolutionary pressures driving these stress-associated NAD+ biosynthesis patterns, we explore the comparative physiology of three other instances of NAD+ biosynthetic alteration: inflammatory activation, pregnancy, and cancer.

Inflammatory Activation

Tryptophan, the first metabolite of the de novo NAD+ biosynthesis pathway, is hydrolyzed by either TDO or IDO. These two enzymes are expressed variably among tissues and show different induction patterns. Detailed reviews of tryptophan-catabolizing enzymes exist,⁴⁰ but, in summary, IDO is widely expressed ubiquitously and typically has low activity. It is induced by proinflammatory cytokines such as tumor necrosis factor-α, interleukin 6, interferon-γ, and prostaglandin E2.^41,42 TDO is expressed primarily in the liver and is not induced by inflammatory cytokines. Therefore, in most tissues, activation of TRP metabolism through the de novo NAD+ biosynthesis pathway is a component of an immune response. This has been well described in inflammatory cells, including macrophages and dendritic cells, but there is evidence that these processes also occur in nonhepatic stromal cells⁴³ and epithelial cells, including renal epithelial cells.⁴⁴

Rather than inciting inflammation, TRP metabolism may correlate with immune tolerance. In inflammatory cells especially, many transcription factors that drive inflammatory activation are nutrient responsive. TRP depletion activates general control nonderepressable 2 (GCN2) stress kinase in dendritic cells and macrophages, which leads to production of anti-inflammatory molecules interleukin 10 and transforming growth factor-β, and suppression of proinflammatory interleukin 12.⁴⁵ Likewise, TRP depletion inhibits mechanistic target of rapamycin complex 1 (mTORC1), a potent driver of inflammation, in HeLa cells.⁴⁶

IDO-induced TRP reduction is accompanied by an increase in downstream metabolites including kynurenine, kynurenic acid, and 3-hydroxyanthranilate (3-HA), and which also play a role in immune tolerance (Fig. 3). Many of these metabolites have been described as aryl hydrocarbon receptor (AhR) ligands.^47,48 AhR is a transcription factor whose activation drives differentiation of naive T cells into regulatory T cells, which are involved in immune tolerance and inflammation resolution⁴⁹ (Table 1). The last committed metabolite of the de novo NAD+ biosynthesis pathway, QA, also accumulates in tissues in periods of inflammation and injury. This has been particularly well described in the kidney^2,13,50–52 and brain.⁵³ The exact effect of QA on inflammation and the mechanism of that effect has yet to be elucidated, although it is believed that QA is an agonist of the N-methyl-D-aspartate (NMDA) receptor.⁵⁴ Little is known about the NMDA receptor in the kidney, but it is expressed in the renal cortex, has been activated and implicated in AKI severity,⁵⁵ and its antagonism has protected against AKI in animal models.⁵⁶

Figure 3.

Targeted accumulation of de novo nicotinamide adenine dinucleotide (NAD+) biosynthesis metabolites. Abbreviations: AFMID, arylformamidase; IDO, indoleamine 2,3-dioxygenase; KMO, kynurenine 3-monooxygenase; KYNU, kynureninase; QPRT, quinolinic acid phosphoribosyltransferase; TDO, tryptophan-2,3-dioxygenase; 3HAO, 3, hydroxyanthranilate 3,4 dioxygenase.

Table 1.

Metabolites of NAD+ Biosynthesis and the Receptors They May Bind

Metabolite	Symbol and Aliases	Structure	Receptor Binding
Tryptophan	L-tryptophan Tryptophane L-tryptophane TRP W
N-formylkynurenine	Formylkynurenine NFK KYNA
Kynurenine	L-kynurenine		Aryl hydrocarbon receptor N-methyl-D-aspartate receptor α-Amino-3-hydroxy-5-methyl-4-isoxazole propionic acid Kainate α7 nicotinic acetylcholine receptors GABA
3-Hydroxykynurenine	3-Hdyroxy-dl-kynurenine Hydroxykynurenine 3-HK 3-OHKY
3- Hydroxyanthranilic acid	3- Hydroxyanthranilate 3-Oxyanthranilic acid 3-HAA
2-amino 3-carboxymuconic-6-semialdehyde
Quinolinic acid	Quinolinate QA QUIN		N-methyl-D-aspartate receptor
Nicotinic acid ribonucleotide	Nicotinate ribonucleotide Nicotinic acid mononucleotide Nicotinate mononucleotide NAMN
Nicotinic acid adenine dinucleotide	Deamido-NAD+ Deamino-NAD+ NAAD NAADN		Two-pore channels (TPCs)
Nicotinic acid	Niacin		Niacin receptor 1 (NIACR 1 or GPR109A) Niacin receptor 2 (NIACR 2 or GPR109B)
Nicotinamide riboside	Nicotinamide ribose N-ribosylnicotinamide Nicotinamide-β-riboside Nicotinamide ribonucleotide Ribosylnicotinamide NR
Nicotinamide mononucleotide	Nicotinamide ribonucleotide NMN
Nicotinamide	Vitamin B3 Niacinamide Nicotinic acid amide Nicotinic amide Vitamin PP

Another study in macrophages corroborated the role of de novo NAD+ metabolites in immune response, although with conflicting findings.¹⁶ Tracing studies have shown that significant macrophage NAD+ comes from TRP and that genetic or pharmacologic blockade of de novo NAD+ biosynthesis depleted NAD+ and impaired macrophage phagocytosis. Lipopolysaccharide (LPS) induced all enzymes of the de novo NAD+ biosynthesis pathway except QPRT, which was suppressed. This led to decreased TRP and accumulation of all other pathway metabolites along with reduced NAD+ (Fig. 3). In this instance, the TRP depletion and accompanied kynurenine metabolite accumulation was associated with expression of proinflammatory cytokines, but reduced monocyte phagocytic activity. Overexpressing QPRT restored NAD+ in LPS-treated cells and restored macrophages’ anti-inflammatory state.¹⁶

Another study showed that activation of the de novo NAD+ biosynthesis pathway versus the NAD+ salvage pathway was associated with an immune tolerance. They did so by stimulating monocytes with high- and low-dose LPS. Low-dose LPS led to NAMPT activation, which led to a proinflammatory state upon secondary LPS exposure. In contrast, high-dose primary stimulation led to de novo NAD+ biosynthesis activation, which resulted in an immune-tolerant phenotype upon secondary LPS exposure.⁵⁷ Both doses stimulated NAD+ biosynthesis, increased NAD+, and increased SIRT1 activity. However, de novo NAD+ biosynthesis induction also increased RelB, a transcription factor that can amplify SIRT1’s transcriptional repression of inflammatory genes such as tumor necrosis factor-4ɑ to induce immune tolerance rather than inflammation.⁵⁷ Although this study did not propose a mechanism whereby de novo NAD+ biosynthesis activated RelB to alter SIRT1 activity, RelB is known to interact with AhR, a receptor that can be modulated by many de novo biosynthesis metabolites.⁵⁸ In this instance, the pathway used to synthesize NAD+ correlated with the way in which NAD+ was used in the immune response. This may represent a critical finding that could elucidate the significance of multiple evolutionarily conserved pathways for NAD+ biosynthesis. Intriguingly, metabolomic and transcriptomic analysis of the stressed kidney show similar patterns as inflammatory cells. Noxious stimuli increase TRP conversion and suppress QPRT, leading to kynurenine precursor accumulation.²

Pregnancy

Pregnancy is another physiologic condition with noted alterations in NAD+ biosynthesis. A 1920 study of pellagra in the United States showed that pellagra was most common among women aged 25 to 45, and that multiparous women were at highest risk.⁵⁹ Regarding TRP metabolism, pregnant women and rats excreted more 3-HA and QA in their urine than nonpregnant subjects.^60,61 Pregnant human beings and rats also excreted more NAD+ consumption products, including NAM and methyl-NAM.^60,62,63 Excretion rates increased throughout pregnancy before peaking close to parturition: around 36 weeks’ gestation in human beings and 15 days’ gestation in rats.⁶⁰ Notably, urine metabolite levels did trend closer to the nonpregnant state after delivery, but not within the first 2 weeks postpartum, implying that it was not strictly fetal contribution to NAD+ metabolism that caused the changes.^60,63

Studies in human beings and rats showed that similar doses of TRP in pregnant individuals led to a much larger increase in NAM and methyl-NAM excretion than in nonpregnant women, implying that pregnancy is associated with more efficient conversion of TRP to NAD +.^60,61 Furthermore, methyl-NAM excretion increased throughout pregnancy and persisted even in the setting of low precursor intake. In many pregnant women, the amount of excreted NAD+ breakdown products exceeded the quantity of precursors ingested, indicating that maternal NA and TRP stores may be used and even depleted during pregnancy to meet NAD+ needs.⁶³ Trials of supplementing pregnant women with NAM did increase excretion of methyl-NAM, but not in equal levels to NAM intake. Therefore, a large portion of the supplemented NAM likely was used for NAD+ metabolism.⁶² Although NA and TRP deficiency severe enough to cause pellagra now is rare in developed countries, a 2002 survey of pregnant women in New Jersey showed that even among subjects taking the recommended perinatal vitamins—which typically include 20 mg NA—evidence of progressive maternal NA deficiency was observed throughout gestation.⁶⁴

Induction of de novo NAD+ biosynthesis during pregnancy may be a component of maternal immunotolerance of the fetus.⁶⁵ Mammalian syncytiotrophoblast cells, which separate fetal and maternal tissues in the placenta, express TRP-consuming IDO, and expression increases throughout gestation.⁶⁶ Treating pregnant mice with the IDO inhibitor, 1-methyl-tryptophan, resulted in loss of all fetuses in a typical semi-allogenic pregnancy whereas fetuses from a syngeneic pregnancy were unaffected.⁶⁶ This effect appeared to be T-cell mediated because dams that were genetically unable to develop lymphocytes did not reject their fetuses even when exposed to 1-methyl-tryptophan.

When analyzing these results in totality, the exact implications are not completely understood. This is, in part, because most studies rely on urine or plasma data without placenta, uterine, or fetal levels to corroborate. We therefore cannot even determine whether the surplus NAD+ byproducts and precursors being excreted represent systemic surplus, hyperutilization, or hypersecretion. Although it does seem that IDO activation and accumulation of kynurenine metabolites are critical to maternal tolerance of the fetus, there is still debate on whether the metabolites themselves or the tryptophan depletion drive the immune phenotype. Mechanistic studies to that effect are needed. Finally, it has been well demonstrated that sufficient NAD+ is critical for normal fetal development,⁶⁷ so maternal adaptations may be suited to meet that need, even at the cost of maternal health.

Cancer

Cancers, even within single organs, are heterogeneous and thus have variation in the way they adapt to synthesize and use NAD+. Indeed, modulations in NAD+ biosynthesis are well-described mechanisms associated with malignant transformation. It is hypothesized that such adaptations may enhance propagation and avoid immune response. Cancer cells shift from oxidative phosphorylation to glycolysis in a phenomenon known as the Warburg effect.^68,69 Although this shift is less efficient for adenosine triphosphate (ATP) production, it confers other survival and propagation benefits. Glycolysis creates building blocks for cell growth including substrates for the pentose phosphate pathway, which is critical for nucleic acid synthesis,⁷⁰ and the serine synthesis pathway, which supports transportation and methylation of histones and DNA via the methionine cycle.^71,72 Finally, glycolysis produces lactate, which acidifies the tissue microenvironment, further enhancing cancer survival.⁷³

NAD+ is required for the glyceraldehyde-3-phosphate dehydrogenase enzyme, which initiates glycolysis; the lactate dehydrogenase enzyme, which completes the final conversion to lactate; and the phosphoglycerate dehydrogenase enzyme, which shuttles glycolysis intermediates toward serine synthesis. The pentose phosphate pathway requires phosphorylated NAD+. Thus, several metabolic alterations associated with malignant transformation depend on enhanced NAD+ production.

There is wide variation in NAD+ biosynthetic alterations in malignant transformation. Because most tissues only use the NAD+ salvage pathway for NAD+ production, it frequently is amplified in cancer, including its rate-limiting enzyme, NAMPT. NAMPT up-regulation has been described in breast cancer, colorectal cancer, thyroid cancer, gastric carcinoma, ovarian cancer, prostate cancer, and malignant gliomas.⁷⁴ Furthermore, chemical NAMPT inhibition has shown promising tumor reduction in experimental models with NAMPT overexpression.^75–77 Unfortunately, translation of these therapies into human use has been complicated by adverse events and only mild cancer improvement.⁷⁸

The PH pathway of NAD+ biosynthesis is altered variably in cancer. Tumors arising from tissues with a highly expressed PH pathway, measured through the rate-limiting enzyme NAPRT, are more likely to overexpress NAPRT after malignant transformation, and even depend on NAPRT for survival.⁷⁹ Studies are ongoing to determine the significance of NAPRT versus NAMPT preference in NAD+ amplification and how that affects cancer outcomes.^80,81

The de novo NAD+ biosynthesis pathway has been less well studied in cancers because most tissues do not express the enzymes of the de novo biosynthesis pathway. However, there is evidence that some malignancies can begin expressing these enzymes as part of the malignant transformation. These adaptations may further optimize NAD+ availability or perhaps augment an anti-inflammatory milieu to promote tumor growth. In published studies brain tumors, bladder carcinomas, melanomas, and hepatocarcinomas overexpressed TDO for TRP catabolism and accumulation of kynurenine metabolites. These metabolites may contribute to immunosuppression via AhR activation.^82,83 TDO inhibition led to appropriate immune rejection of tumor cells.⁸³ Another study showed that many human tumor cell lines constitutively express IDO,⁸⁴ and inhibiting IDO had an antitumorigenic effect.^85,86

Additional evidence has shown that these metabolites may pass between cell types in patterns that promote malignant transformation. For example, gliomas have been shown to express TDO and generate kynurenine, which may assist in immune escape. However, both astrocytoma and glioblastoma cells also have been shown to accumulate QA, a downstream metabolite from kynurenine.⁸⁷ Interestingly, glioma cells do not express 3-HA oxygenase, the enzyme required to convert kynurenine to QA. Therefore, it was hypothesized that QA may come from infiltrating monocytes that do express 3-HA oxygenase. Furthermore, malignant gliomas expressed QPRT to convert QA to NAD+, a capability not possessed by non-neoplastic astrocytes.⁸⁷ The utilization of the de novo NAD+ biosynthesis pathway was accompanied by a near-elimination of the PH pathway expression.⁸⁷ Although this same study showed that QPRT expression was associated with worse glioma grade and disease severity, the physiological significance of using QA versus NA for NAD+ biosynthesis is not yet understood. Perhaps it is a compensatory shift to overcome NA depletion in a rapidly growing cell population. However, given the known association of QA with many other neurologic conditions and the physiologic relevance of other kynurenine pathway metabolites, further study is warranted to understand the role of QA accumulation in malignancy beyond altered NAD+ production.

RETURNING TO THE KIDNEY

Surveying nonkidney settings of altered NAD+ metabolism showed several patterns. First, there were examples showing that switching from one NAD+ biosynthetic pathway to another may have changed cellular response to inflammation and growth inhibition. Therefore, the evolution of three distinct pathways of NAD+ production may not merely reflect the biologic significance of NAD +. Instead, NAD+ biosynthetic pathway switching within renal parenchyma may enable cells to coordinate interactions with the immune system or exert control over other biosynthetic processes. Second, multiple contexts offer examples of simultaneous induction of IDO/TDO and suppression of QPRT, two changes that lead to accumulation of kynurenine metabolites. This phenomenon of triggered kynurenine metabolite accumulation, whatever its significance, also appears to be a conserved feature of the renal stress response.^{2,4,6,12,13,88}

The emerging literature in the kidney and elsewhere therefore poses several questions for future consideration. What is the role of NAD+ in affecting immune cell activation in AKI? How much of the NAD+ metabolite changes attributed to AKI are from the infiltrated immune cells, and how much are from native kidney cells? What kind of crosstalk between cells influences NAD+ biosynthesis? What are the stimuli that lead to NAD+ biosynthesis pathway alterations in AKI? What is the significance of kynurenine pathway metabolite accumulation and TRP depletion in the kidney? How does NAD+ augmentation affect pathway switching, intermediary metabolite accumulation, and downstream effects of metabolite accumulation? We need to untangle the effect of each accumulated metabolite and learn the mechanisms by which those effects take place. It has been shown that NAD+ augmentation may be beneficial against acute stress, but it also is becoming clear that NAD+ deficiency is not the only negative effect of biosynthesis alterations. Therefore, we must ask if NAD+ supplementation is merely replacing depleted NAD+ or if its effects are more broadly altering pathway flux and metabolite accumulation in ways that may provide benefit.

In summary, renal NAD+ biosynthesis alteration is a clear component of AKI and harnessing these alterations may offer therapeutic promise. However, there is much more to learn. Although the kidney is unique in the way it synthesizes and uses NAD+, multiple different cell types and physiological phenomena also show specialized NAD+ utilization. Stepping back to assess the similarities and differences as well as the physiologic triggers and outcomes between the different scenarios may be key to understanding what is happening in the acutely stressed kidney.