Ceramides and Acute Kidney Injury

Summary

Altered lipid metabolism is a characteristic feature and potential driving factor of acute kidney injury (AKI). Of the lipids that accumulate in injured renal tissues, ceramides are potent regulators of metabolism and cell fate. Up-regulation of ceramide synthesis is a common feature shared across several AKI etiologies in vitro and in vivo. Furthermore, ceramide accumulation is an early event in the natural history of AKI that precedes cell death and organ dysfunction. Emerging evidence suggests that inhibition of ceramide accumulation may improve renal outcomes in several models of AKI. This review examines the landscape of ceramide metabolism and regulation in the healthy and injured kidney. Furthermore, we discuss the body of literature regarding ceramides as therapeutic targets for AKI and consider potential mechanisms by which ceramides drive kidney pathogenesis.

The prevalence and costs associated with acute kidney injury (AKI) are increasing in the United States and global populations.^1,2 Despite a significant clinical need, no therapies are currently available to prevent or treat AKI or its transition to chronic kidney disease. Development and validation of candidate therapeutic strategies will require a better understanding of the primary effectors and molecular processes necessary for AKI pathogenesis. Increasing evidence suggests that the proximal tubular epithelium is the primary site affected during AKI and that tubule-specific molecular insults are sufficient to induce kidney pathology. In particular, the form and function of the proximal tubule is tightly coupled to its metabolism. Because of the high energy demand and reliance on aerobic oxidation of fatty acids required to support active solute reabsorption, the proximal tubule is especially vulnerable to metabolic injury. Altered lipid metabolism, characterized by mitochondrial impairments and lipid accumulation, is a shared feature common to diverse AKI etiologies.

Relatively little attention has been paid to the lipotoxic effects of specific lipid species that accumulate in the injured kidney. Nonetheless, bioactive ceramides are candidate mediators of tubular insult and AKI. They are the central molecule of the diverse sphingolipid family. Despite their low abundance, modest fluctuations in intracellular ceramides elicit dramatic effects on cellular metabolism and survival. Evidence collected over the past 25 years has indicated that the increase in renal ceramide concentrations is an early and acute event across numerous models of kidney injury. Furthermore, ceramide-lowering interventions show promise in preventing the onset and severity of AKI. Herein, we summarize the current state of knowledge regarding sphingolipid metabolism with a focus on ceramides in the healthy and injured kidney.

CERAMIDE SYNTHESIS

Ceramides are produced from one of three predominant pathways (Fig. 1). The first mechanism for ceramide biosynthesis is the de novo pathway residing within the endoplasmic reticulum (ER), which commences with the condensation of palmitate and serine via the serine palmitoyl transferase heterotrimer to form 3-ketosphinganine.³ This product is quickly reduced by 3-ketosphingosine reductase to form the characteristic sphingoid backbone dihydrosphingosine. Next, a family of six dihydroceramide synthase enzymes (CERS1-6) incorporate a second acyl group of varying lengths (eg, 16-26 carbons) to the backbone to form the dihydroceramides.⁴ Ultimately, the addition of a 4,5-trans-double bond by the dihydroceramide desaturases (DES1-2) converts a dihydroceramide to a ceramide.

Figure 1.

Ceramide synthesis. Ceramides are produced via de novo biosynthesis, sphingomyelin hydrolysis, or salvage of sphingoid bases. The ubiquitous de novo pathway is characterized by the transfer of serine to palmitoyl CoA, producing a sphingoid backbone that subsequently is acylated and desaturated to form ceramide. Alternatively, cleavage of a choline headgroup by a family of sphingomyelinase enzymes rapidly converts sphingomyelin to ceramide. Lastly, re-acylation of liberated (dihydro)sphingosine by the (dihydro)ceramide synthase enzymes regenerates ceramides in the salvage pathway. The length of acyl chain incorporated into a ceramide molecule is dependent on the substrate specificity of the (dihydro)ceramide synthases. Abbreviations: CDase, ceramidase; CERS, (dihydro)ceramide synthase; DAG, diacylglycerol; DES, dihydroceramide desaturase; FFA, free fatty acid; SK, sphingosine kinase; PC, phosphatidylcholine; R, accessory acyl chain; SK, sphingosine kinase; SM, sphingomyelin; SMase, sphingomyelinase; SMS, sphingomyelin synthase; SPT, serine palmitoyltransferase; S1P, sphingosine 1-phosphate; 3KSR, 3-ketosphinganine reductase. Figure was created with BioRender.com.

From the ER, dihydroceramides and ceramides can be shuttled to the Golgi apparatus or mitochondria for the formation of more complex sphingolipids, such as (dihydro)sphingomyelin, via the addition of various head-groups (eg, choline, phosphate, carbohydrate moieties) to the first-position oxygen molecule.⁵ Hydrolysis of the choline headgroup from sphingomyelin (SM), mediated by a family of sphingomyelinases (SMase), is the second mechanism contributing to ceramide synthesis (Fig. 1).⁶

Lastly, dihydroceramides and ceramides are degraded by the removal of the variable acyl chain to form (dihydro)sphingosine and a free fatty acid via a group of ceramidases (CDase)⁷ or the ligand-gated ceramidase activity of the adiponectin receptors.⁸ The liberated sphingoid backbone can be phosphorylated by sphingosine kinase to form (dihydro)sphingosine-1-phosphate, which has unique bioactivity relevant to kidney diseases, and was recently reviewed elsewhere.^9,10 Re-acylation of (dihydro)sphingosine by CERS to form (dihydro)ceramide is the third and final mechanism of ceramide production, termed the salvage pathway (Fig. 1).¹¹

REGULATION OF CERAMIDE METABOLISM

Not all ceramides are created equal, and their diversity in bioactivity and tissue/intracellular distribution is influenced by a myriad of factors. First, the CERS enzymes vary greatly in tissue expression pattern and substrate specificity. Ceramides containing a long-chain (LC) 16-carbon accessory group (ie, Cerd18:1/16:0) produced by CERS5 and CERS6 are implicated as primary culprits of ceramide-driven pathology.^12–14 Alternatively, very-long-chain (VLC) ceramides produced by CERS2 and CERS3, such as Cerd18:1/24:0, are broadly considered benign or beneficial.^15,16 Thus, aside from fluctuations in total ceramide levels, an increase in the ratio of LC to VLC ceramides is a meaningful indicator of ceramide-related disease risk and prognosis.¹⁷

Sphingolipid metabolism is highly organized and compartmentalized within the cell because of the pH optima of ceramide-metabolizing enzymes. Enzymatic players favoring a neutral pH, such as neutral SMase or neutral CDase, are capable of modulating the sphingolipid composition of the plasma membrane; whereas acid SMase or acid CDase primarily metabolize sphingolipids within the lysosome. As such, regulation of cellular processes is likely modulated by altered sphingolipid metabolism within specific organelles, cellular compartments, or membrane interfaces. For example, ceramide-metabolizing enzymes have been isolated from purified inner- and outer-mitochondrial membranes, as well as mitochondrial-associated membranes interfacing with the ER.^18–23 Much remains to be discovered regarding the nature of ceramide metabolism at subcellular resolution. These upcoming investigations will yield valuable insights regarding ceramide-mediated regulation of physiological and pathophysiological cellular biology.

In addition to the unique tissue and subcellular locales of sphingolipid-related enzymes and their resulting lipid products, we highlight the abundance of physiological factors that modulate sphingolipid-related enzyme activity. De novo biosynthesis of ceramides is activated under conditions of nutrient overload owing to the increased availability of palmitate and serine.²⁴ In addition, ceramide metabolism is responsive to hypoxia and reoxygenation. Because the conversion of dihydroceramide to ceramide requires an oxygen acceptor molecule, DES1 activity is acutely regulated by oxygen availability,^25,26 and several reports have documented dihydroceramide accumulation and/or ceramide depletion with hypoxia or ischemia.^25,27,28 As such, restoration of DES1 activity and the resulting acute spike in ceramides produced after reoxygenation may be a primary mechanism driving ischemia-reperfusion injury. Ceramides are established secondary messengers produced in response to a multitude of stress stimuli, including tumor necrosis factor α, interleukin 1β, lipopolysaccharide, and oxidative stress, among others (reviewed in Nikolova-Karakashian and Rozenova²⁹). Lastly, the ligand-gated ceramidase activity of the adiponectin receptors may explain the protective effects of adiponectin observed in kidney injury and disease models.^30–34 Importantly, these mechanisms for stimulation of ceramide production or degradation are highly relevant to the pathophysiology of AKI and position ceramides as a common denominator downstream of a broad range of kidney insults.

CERAMIDE EXPRESSION IN HEALTHY KIDNEYS

Ceramides are expressed throughout the kidney, albeit at much lower concentrations than the more abundant glycerolipid species (eg, triacylglycerol, diacylglycerol, phospholipids). An early report measuring ceramides in kidney tissue by high-throughput liquid chromatography mass spectrometry in 1999 by Kalhorn et al³⁵ documented that the total ceramide pool in kidney cortex comprised 20% Cer(d18:1/16:0) and 70% Cer(d18:1/24:0). In subsequent studies, this trend has remained mostly consistent, with Cer(d18:1/16:0) and Cer(d18:1/24:0) being the most abundant species in the kidney at relatively equal concentrations, followed by Cer(d18:1/24:1) (unpublished data and Dupre et al,³⁶ Eckes et al,³⁷ Laviad et al³⁸). Interestingly, expression profiling of CERS in mouse kidney tissue suggests that Cers2 is the most abundantly expressed Cers messenger RNA, accounting for approximately 85% of total Cers expression, whereas Cers6 accounts for just 5%.³⁸ Thus, Cers message expression and kidney ceramide acyl-chain composition are not entirely proportional. Very little has been done to characterize CERS protein expression in kidney tissue, and the correlation of Cers expression to ceramide also could be affected by differential ceramide turnover according to acyl-chain length.³⁹

Few studies have reported the spatial distribution of renal ceramides. Eckes et al³⁷ profiled ceramide concentrations in the cortex and medulla of healthy human kidney tissue collected via nephrectomy. Notably, the overall abundance of ceramides was two-fold higher in the cortex versus medulla. This difference was attributable mainly to decreases in medulla VLC ceramide species (ie, Cer[d18:1/24:0] and Cer[d18:1/24:1]), which were nearly two- and three-fold lower in the medulla versus cortex, respectively.³⁷ Mass spectrometry imaging of mouse and rat kidneys also has been achieved. In a mouse model of Farber’s disease, loss of acid CDase activity led to global increases in all kidney ceramide species.⁴⁰ Spatial lipidomic profiling of Farber’s kidneys showed enrichment of Cer(d18:1/16:0) and Cer(d18:1/24:1) in the cortex, with Cer(d18:1/18:0) expression localized to the medulla.⁴⁰ In a separate study, mass spectrometry imaging of healthy rat kidneys showed dispersal of Cer(d18:1/16:0) throughout the cortex and medulla, with enrichment in the corticomedullary region.⁴¹ Cer(d18:1/24:0) was localized to the cortex, Cer(d18:1/22:0) to the medulla, and Cer(d18:1/20:0) to the corticomedullary junction.⁴¹ Thus, the spatial distribution of ceramides throughout the kidney deserves further investigation and resolution.

Transcriptomics of kidney single cells and/or nuclei are becoming widely published and available for investigation of differential expression of genes across renal cell types.⁴² Notably, a Kidney Tissue Atlas recently was published by the Kidney Precision Medicine Project, including single-nucleus RNA sequencing (snRNAseq) of 13 healthy human reference samples (https://atlas.kpmp.org and Hansen et al⁴³). We probed these data to visualize expression of 29 ceramide-related genes across 25 distinct cell clusters spanning the nephron (Fig. 2). These data showed surprising heterogeneity in ceramide-related gene expression across cell populations (eg, high CERS6 expression in the podocyte, high SPTLC2 expression in the loop of Henle). In addition, we were surprised that CERS2 expression was relatively low across cell clusters, in disagreement with reports of high renal CERS2 expression by reverse-transcription polymerase chain reaction.^4,38 This discrepancy could be owing to the overall scaling of a modestly abundant message expressed within the predominant cell type present in whole tissue (ie, proximal tubule), or an artifact of gene length-associated detection bias resulting in poor coverage of short (<15 kb) genes reported with snRNAseq versus single-cell RNAseq.^44–47 Specifically, the CERS2 gene is relatively short (14 kb) and is detected at higher expression levels throughout the tubular epithelium in human⁴³ and mouse⁴⁸ single-cell RNAseq or mouse spatial transcriptomics⁴⁹ data sets. Accordingly, transcriptional and protein expression of acid SMase and neutral SMase 1 encoded by SMPD1 and SMPD2, respectively are reportedly high in the kidney⁵⁰; yet these short gene products (<5 kb) are poorly represented in the snRNA-seq data set. This point again highlights the need to resolve spatial distribution of ceramide species and their association with cellular or spatial ceramide-related gene and protein expression.

Figure 2.

Expression of ceramide-related genes throughout the nephron. Ceramide-related gene expression throughout the nephron cell types and segments was quantified via single-nucleus RNA sequencing in healthy human kidneys (n = 13) and accessed via the Kidney Precision Medicine Project Kidney Tissue Atlas.⁴³ Gene expression is characterized by the percentage of cells within each identified cluster expressing the target (dot size) and the average relative expression of the cell cluster (dot hue). Abbreviations: ATL, ascending thin limb cell; CCD-IC-A, cortical collecting duct intercalated cell type A; CCD-PC, cortical collecting duct principal cell; CNT, connecting tubule cell; CNT-IC-A, connecting tubule intercalated cell type A; CNT-PC, connecting tubule principal cell; C-TAL, cortical thick ascending limb cell; DCT1, distal convoluted tubule cell type 1; DCT2, distal convoluted tubule cell type 2; DTL1, descending thin limb cell type 1; DTL2, descending thin limb cell type 2; DTL3, descending thin limb cell type 3; IC-B, intercalated cell type B; IMCD, inner medullary collecting duct cell; MC, mesangial cell; MD, macula densa cell; M-TAL, medullary thick ascending limb cell; OMCD-IC-A; outer medullary collecting duct intercalated cell type A; OMCD-PC, outer medullary collecting duct principal cell; PapE, papillary tip epithelial cell; PEC, parietal epithelial cell; POD, podocyte; PT-S1, proximal tubule epithelial cell segment 1; PT-S2, proximal tubule epithelial cell segment 2; PT-S3, proximal tubule epithelial cell segment 3. Figure was created based on data generated by the Kidney Precision Medicine Project (KPMP): DK114886, DK114861, DK114866, DK114870, DK114908, DK114915, DK114926, DK114907, DK114920, DK114923, DK114933, and DK114937 (https://www.kpmp.org; data were downloaded on July 1, 2022).

ALTERED METABOLISM OF CERAMIDES IN AKI

Generation of renal ceramides is a shared feature across several AKI etiologies, and numerous reports have detailed increases of ceramides in cultured tubular cells (Table 1) or rodent kidney tissue (Table 2) after acute insult. In vitro, challenging renal tubular cells (RTC) with a variety of nephrotoxins (ie, cisplatin,⁵¹ radiocontrast media,⁵² cadmium,⁵³ nickel,⁵⁴ oxalates,⁵⁵ polyene antibiotics,⁵⁶ and fluorinated anesthetics⁵⁷), oxidizing agents (ie, ferrous iron⁵⁸ and hydrogen peroxide⁵⁹), chemical hypoxia,⁶⁰ or hypoxia/reoxygenation⁶¹ elicits significant increases in ceramide levels. Dupre et al³⁶ reported increases in kidney cortex ceramides in vivo 72 hours after intraperitoneal cisplatin administration (25 mg/kg) in male C57BL/6J mice. These findings were supported by findings in male Sprague Dawley rats, wherein cortical and medullary Cer(d18:1/16:0) was increased in kidneys 7 days after cisplatin injection (2.5-10 mg/kg).⁶² In cohorts of rhabdomyolysis-associated AKI, cortex ceramides were increased markedly at 2, 18, and 24 hours after intramuscular glycerol injection (50% glycerol solution, 10 mL/kg) in male CD-1 mice^35,63 and at 1 day postinjection in male Sprague Dawley rats.³⁶ CCl₄ nephrotoxicity was associated with increased renal ceramides 24 hours or 1 week after toxin administration, respectively, in male Wistar rats (4 mL/kg; 1:1, v/v in mineral oil)⁶⁴ or male ICR mice (2 × 2 mL/kg; 1:1, v/v in peanut oil).⁶⁵ Renal ceramide generation also has been reported in postrenal AKI 14 days after unilateral ureteral obstruction in adult Wistar rats⁶⁶ and neonatal Sprague Dawley rats,⁶⁷ although this observation recently was contested by Eckes et al.³⁷ Perhaps the most robustly characterized in vivo response of ceramides to AKI in vivo is with ischemia/reperfusion (I/R) injury. Specifically, assessment of ceramides after a 45-minute unilateral occlusion of the left renal artery and vein in male CD-1 mice elicited acute increases in cortex ceramides as early as 30 minutes^27,63 and for as long as 18 hours after organ reperfusion.³⁵ Additional studies have reported increases in cortex ceramides in male C57BL/6J mice 24 hours after bilateral I/R injury.^36,68

Table 1.

Response of Ceramides to In Vitro Models of Acute Kidney Injury in Renal Tubule Cells

Stressor	Ceramides	RTC Line/Source	Mechanism
Hypoxia27,60,61,71	↑	HK-2, LLC-PK1, isolated tubules	↑ De novo
H/R61	↑	NRK-52E	↑ De novo
Oxidant
Fe58	↑	HK-2	↑ De novo or ↓ CDase
H2O259	↑	LLC-PK1	↑ De novo
Toxin
Cisplatin51	↑	Baby mouse kidney cells	↑ De novo
Fluorinated anesthetic57	↑	HK-2, isolated tubules	UN
Radiocontrast media52	↑	LLC-PK1	↑ De novo
Cadmium53	↑	WKPT-0293 C1.2	↑ De novo
Nickel54	↑	WKPT-0293 C1.2	↑ De novo
Oxalate55,73	↑	LLC-PK1, NRK-52E, MDCK	↑ SMase
Myohemoglobin27	↔	Isolated tubules	UN
Polyene antibiotic56	↑	Isolated tubules	UN
Irradiation51	↑	Baby mouse kidney cells	↑ De novo
Arachidonic acid58	↑	HK-2	↑ SMase ↓ CDase

Abbreviations: CDase, ceramidase; HK-2, human kidney 2; H/R, hypoxia/reoxygenation; LLC-PK1, Lilly laboratories culture-porcine kidney 1; MDCK, Madin-Darby canine kidney; NRK-52E, normal rat kidney-52E; PLA₂, phospholipase 2; RTC, renal tubule cell; SMase, sphingomyelinase; UN, unknown; WKPT, Wistar–Kyoto rat proximal tubule.

Table 2.

Response of Kidney Ceramides to In Vivo Models of Acute Kidney Injury

Kidney Insult	Ceramides	Species: Strain	Kidney Tissue	Mechanism
Ischemia	↓	Mouse: CD-127	Cortex	↓SMase
I/R	↑	Mouse: CD-127,35,63 Mouse: C57BL/6J36,68	Cortex	↑ De novo ↑↓SMasea
Rhabdomyolysis	↑	Mouse: CD-135,63 Rat: Sprague Dawley36	Cortex	↑ De novo ↑↓SMaseb
Cisplatin	↑	Mouse: C57BL/6J36 Rat: Sprague Dawley62	Cortex, medulla	↑ De novo ↑SMase
UUO	↑↓c	Mouse: C57BL/6J37 Rat: Wistar66 Rat: Sprague Dawley67	Cortex, whole	UN
αGBM antibody	↑	Rat: Sprague Dawley63	Cortex	↑ SMase
Fluorinated anesthetic	↑	Mouse: CD-157	Cortex	UN
CCl4	↑	Mouse: ICR65 Rat: Wistar64	Whole	↑ SMase

Abbreviations: I/R, ischemia/reperfusion; αGBM, anti–glomerular basement membrane; SMase, sphingomyelinase; UN, unknown; UUO, unilateral ureteral obstruction.

^aDecreased acid sphingomyelinase and neutral sphingomyelinase (nSMase) activity according to Zager et al,⁶³ and increased acid sphingomyelinase activity reported by Dupre et al³⁶ after ischemia/reperfusion.

^bDecreased acid sphingomyelinase and nSMase activity according to Zager et al,⁶³ and increased acid sphingomyelinase activity reported by Dupre et al³⁶ after myohemoglobinuria.

^cAfter unilateral ureteral obstruction, ceramides reportedly decreased according to Eckes et al,³⁷ and increased according to Verdoorn et al⁶⁶ and Malik et al.⁶⁷

Aside from observations of ceramide accumulation in AKI, several studies have meaningfully ameliorated or worsened injury to kidneys or RTCs with loss- or gain-of-ceramide interventions, respectively (Table 3).^69,70 Notably, the success of ceramide-lowering treatments in improving kidney outcomes coordinates with the proposed mechanism of ceramide accumulation associated with the model (Tables 1 and 2). In vitro, inhibition of serine palmitoyl transferase or CERS activity via myriocin or fumonisin B, respectively, prevents cell apoptosis related to treatment with hypoxia, hypoxia/reoxygenation, hydrogen peroxide, cadmium, radiocontrast media, ultraviolet-C (UV-C) irradiation, or cisplatin.^{51–53,59–61} Conversely, inhibition of SM hydrolysis does not prevent ceramide accumulation or cell death in RTCs after radiocontrast media exposure—an insult for which de novo ceramide synthesis is most indicated.⁵² In mice, co-inhibition of de novo synthesis and SM hydrolysis via myriocin and amitriptyline, respectively, successfully ameliorated cisplatin-induced AKI,³⁶ although inhibition of each pathway was not tested individually. Notably, pharmacologic inhibition of glucosylceramide synthesis via C10 administration exacerbated ceramide accumulation and worsened AKI severity after cisplatin treatment.³⁶ Together, these studies warrant further investigation of ceramide as a potential therapeutic target for AKI.

Table 3.

Effects of Gain- or Loss-of-Ceramide Interventions in AKI

Intervention	Target/Activity	Ceramide	Model	System	Outcome
Fumonisin B1	CERS / ↓	↓	H/R	In vitro, NRK-52E cells	Decreased apoptosis61
		↓	Hypoxia	In vitro, LLC-PK1 cells	Decreased apoptosis60
		↓	H2O2	In vitro, LLC-PK1 cells	Decreased apoptosis59
		↓	Cadmium	In vitro, WKPT-0293 C1.2 cells	Decreased apoptosis53
		↓	Radiocontrast media	In vitro, LLC-PK1 cells	Decreased apoptosis52
		No change	Fluorinated anesthetic	In vivo, CD-1 mice	No protection57
Myriocin	SPT / ↓	↓	UV-C light	In vitro, baby mouse kidney cells	Decreased apoptosis51
		↓	Cisplatin	In vitro, baby mouse kidney cells	Decreased apoptosis51
Myriocin + amitriptyline	SPT / ↓ aSMase / ↓	↓	Cisplatin	In vivo, C57BL/6J mice	Decreased AKI36
L-cycloserine	SPT / ↓	UN	Radiocontrast media	In vitro, LLC-PK1 cells	Decreased apoptosis52
D609	SMase / ↓	No change	Radiocontrast media	In vitro, LLC-PK1 cells	No protection52
C10	GCS / ↓	↑	Cisplatin	In vivo, C57BL/6J mice	Worsened AKI36
Exogenous SMase	SMase / ↑	↑	CAD	In vitro, HK-2 cells	Increased apoptosis69
		↑	Fe	In vitro, NRK-52E, MDCK cells	Increased apoptosis56
nCDase KO	nCDase / ↓	UN	Cisplatin	In vivo, male C57BL/6J mice	Decreased AKI70

Abbreviations: AKI, acute kidney injury; aSMase, acid sphingomyelinase; CAD, calcium ionophore/antimycin A/2-deoxyglucose; CERS, dihydroceramide synthase; GCS, glucosylceramide synthase; HK-2, human kidney 2; H/R, hypoxia/reoxygenation; KO, knockout; LLC-PK1, Lilly Laboratories cell porcine kidney 1; MDCK, Mmadin-Ddarby canine kidney; nCDase, neutral ceramidase; NRK-52E, normal rat kidney 52E; SMase, sphingomyelinase; SPT, serine palmitoyltransferase; UN, unknown; UV-C, ultraviolet-C; WKPT, Wsytar-Kyoto rat proximal tubule.

POTENTIAL MECHANISMS FOR CERAMIDE ACCUMULATION IN AKI

Despite these findings, our understanding of the processes underlying ceramide accumulation in AKI remains ambiguous. Indeed, the mechanisms by which ceramides are generated are likely multifaceted and variable across origins of kidney insult. The primary processes implicated in ceramide production after AKI are as follows: (1) increased de novo synthesis, (2) increased SM hydrolysis, or (3) decreased ceramide degradation, which have been investigated in multiple contexts of AKI with targeted enzyme inhibitors or activity assays (Tables 1 and 2). Yet, inconsistencies in reported mechanisms exist, even within the same AKI models. For example, Zager et al⁶³ attributed kidney cortex ceramide accumulation after I/R injury to increased de novo ceramide production or decreased CDase activity, as measured acid and neutral SMase activity was consistently reduced during ischemia and for 24 hours after reperfusion. On the other hand, Dupre et al³⁶ observed increases in the activity of both acid SMase and CERS producing LC ceramides 24 hours after reperfusion. Furthermore, Zager et al²⁷ reported a 33% reduction in kidney cortex ceramide concentrations after a 45-minute ischemia period; yet, ex vivo ceramides in isolated tubules subjected to 20 minutes of hypoxia were 20% higher than control samples. Accordingly, renal tubule cells cultured under chemically or physically hypoxic conditions also accumulate ceramides,^60,61,71 with further increases elicited by reoxygenation⁶¹ (Table 1). Aside from potential effects caused by differences in mouse strain, ischemia time, and assay conditions, these discrepancies present several unanswered questions regarding the mechanisms of ceramide generation in AKI.

First, it is yet to be determined which cell type or types accrue ceramides during AKI. Logically, as ceramides increase with various stimuli in cultured renal tubule cells, it is likely that tubule-derived ceramides play some autonomous role in AKI; however, differences in ceramide responses to kidney insults in vivo versus ex vivo and in vitro suggest further complexity. Advances in single-cell and spatial omics studies will advance our understanding of the interplay of sphingolipid metabolism across renal cell populations and regions in response to AKI. For example, snRNAseq of cryopreserved male mouse kidneys after I/R injury showed marked increases in proximal tubule Sptlc2 expression as early as 4 hours after reperfusion, with sustained increased expression in injured proximal tubule clusters for as long as 2 weeks postinjury (http://humphreyslab.com/SingleCell and Kirita et al⁷²).

Second, many initial studies of ceramides and AKI only reported changes in total ceramides. Early methods of ceramide quantification involved a diacylglycerol kinase assay followed by thin-layer chromatography, which is incapable of distinguishing ceramides by their chain length. Current standard methods use mass spectrometry to quantify individual ceramide species and resolve additional intricacies (eg, distinguishing the single degree of unsaturation differentiating a dihydroceramide from a ceramide). Whether specific species are responsible for ceramide-driven pathology in AKI remains unanswered; however, several publications have reported that despite increases in ceramides across all accessory chain lengths, enrichment of LC ceramides is supraproportional relative to VLC ceramides.^35,36 Thus, increases in LC ceramides, such as Cer(d18:1/16:0), could be especially deleterious to kidney function and health, warranting further investigation into the regulation and manipulation of specific ceramide species production in AKI.

A final area of investigation will be to resolve the natural history of ceramide accumulation and action in AKI in settings wherein several mechanisms for ceramide generation may be concerned. For instance, the reported processes implicated in ceramide synthesis after I/R injury are broad and encompass transcriptional up-regulation of Sptlc2^49,68,72; increased activity of LC ceramide production by CERS^35,36; DES1 inactivity/rebound during I/R²⁷; transcriptional down-regulation of Asah2^49,68,72; and up-regulation of acid SMase activity.³⁶ Because of the breadth of physiological stress stimuli expressed during AKI, which are known to modulate ceramide metabolism, these findings are not entirely unexpected. Yet, each individual mechanism driving ceramide production may have unique implications for the following: (1) cellular fate and function, (2) short- versus long-term outcomes after AKI, and (3) potential for therapeutic targeting. Regarding the first two points, strikingly rapid responses of ceramides to insult (ie, on the order of seconds-minutes) attributable to both acute activation of de novo ceramide biosynthesis or SM hydrolysis have been observed before apoptosis in cultured RTCs^59,60,73 and other cell types.^74–79 Yet, much less is known regarding the effects of sustained alterations of the ceramide pool after AKI on cellular metabolism, proliferation, differentiation, and long-term kidney function. Elucidating the time course and associated consequences of ceramide-driven mechanisms in vivo will show novel insights into the pathophysiology and therapeutic development for AKI.

PROPOSED APOPTOTIC AND METABOLIC ACTIONS OF CERAMIDE DRIVING AKI

Although ceramides play a well-documented role in membrane structure and organization, we believe that their pleiotropic effects on cellular metabolism and survival serve an evolutionarily conserved purpose of sensing and responding to increasing intracellular free fatty acid levels. The injured kidney is subject to damage related to fatty acid (FA) overload,^80–83 secondary to both increased tubular uptake and impaired FA oxidation. This imbalance in lipid availability and utilization provokes a shift toward increased storage of FAs as inert triglyceride, which is likely not pathogenic.^84,85 However, insufficient storage or oxidation of amphipathic free FAs subjects cells to increased risk of membrane lysis and tissue injury. In response, modest increases in ceramides elicit a coordinated metabolic program to alleviate the FA burden by 1) enhancing FA esterification and storage as triglyceride and 2) decreasing mitochondrial efficiency, requiring consumption of a larger number of FA substrates to maintain the mitochondrial membrane potential (reviewed by Nicholson et al⁸⁶). Further increases in ceramide levels stimulate a pivot toward programmed cell death and fibrosis processes to minimize overall damage to the tissue or organism caused by uncontrolled cell lysis.⁸⁶ The actions of ceramide to promote lipid accumulation and storage may not be pathogenic per se in the setting of kidney injury. Instead, more evidence suggests that mechanisms of ceramides inciting mitochondrial dysfunction and cell death may be influential features of AKI.

The majority of mechanistic studies of ceramides in AKI have detailed the ceramide-dependent initiation of apoptosis after insult (Table 1). Specifically, ceramides are involved in the intrinsic apoptotic mechanisms governed by the mitochondria. Endogenous or exogenous ceramides govern membrane platforms to initiate mitochondrial outer membrane permeabilization in concert with members of the Bcl-2 family.^51,87,88 Alternatively, ceramides themselves are capable of forming large and stable channels in the mitochondrial membrane allowing release of cytochrome c.^89,90 Notably, only Cer(d18:1/16:0) or exogenous short-chain (eg, Cer[d18:1/2:0]) ceramides are capable of forming stable channels. Furthermore, these structures are disrupted by low concentrations of dihydroceramide or intercalation of VLC ceramides.^15,91 Again, these phenomena highlight the incredible diversity in ceramide bioactivity modulated by the length of the ceramide accessory acyl chain or the presence of a trans-4,5-double bond distinguishing ceramides from dihydroceramides.

Aside from facilitating mitochondrial outer membrane permeabilization, ceramides in the mitochondria exert potent effects on the activity of electron transport within the respiratory chain. Ceramides, in particular Cer (d18:1/16:0), decrease complex I, III, and IV activity; deplete mitochondrial adenosine triphosphate production; and induce mitochondrial depolarization and reactive oxygen species production.^{12,14,22,92–95} Yet again, these effects are not recapitulated with dihydroceramide treatment²² or induction of VLC ceramides.^12,14,96 The mechanism by which ceramides acutely impair respiratory chain complex activity is not fully resolved; however, treating isolated mitochondria with low doses of Cer(d18:1/16:0) elicits near-immediate changes in oxygen consumption, suggesting that ceramides impact inner-mitochondrial membrane composition to modulate respiratory chain activity or interact directly with the proteins themselves.^13,97,98 On a longer time-scale, CERS6-derived ceramides interact with the mitochondrial fission factor to initiate mitochondrial fission and fragmentation.¹³ Because of the high metabolic demand and oxidative phosphorylation activity of the kidney, particularly within the proximal tubule, ceramide-mediated mitochondrial impairment is highly relevant to AKI pathophysiology. Yet, at the time of this review, no investigations have been conducted to characterize ceramide-dependent impairments in renal mitochondrial function and respiration during or after AKI.

CONCLUSIONS

Thus far, the frontier of research exploring the metabolism and mechanisms of ceramides within the kidney remains minimally explored. Preliminary evidence points to complex cellular and spatial organization regulating ceramide metabolism in the kidney. Yet, significant challenges in the fields of sphingolipid and renal metabolism have stalled scientific advancements in this area. For example, progress in spatial and single-cellular proteomic and lipidomic profiling has lagged behind transcriptomic methods. Development and optimization of these technologies will allow for in-depth characterization of ceramide distribution within healthy kidneys and their relation to injured cell signatures. Furthermore, many proposed actions of ceramides are likely related to post-translational, rather than transcriptional, mechanisms. Methodologies to probe for lipid-protein interactions on the molecular level and within subcellular compartments will provide valuable context to ceramide-driven pathology in AKI. Lastly, interventional studies using pharmacologic inhibitors of ceramide-metabolizing enzymes often are complicated by direct renal toxicity of the compound (eg, fumonisin B1, amitriptyline) or its dependence on renal clearance (eg, cycloserine). Development of more specific and potent inhibitors will rely on ongoing characterization of target enzymes’ high-resolution structural information. Lastly, genetic manipulation of ceramide metabolism within the kidney has been minimally exercised. Because of the established roles of sphingolipids in development, these studies will require inducible transgenic manipulations of ceramides within specific kidney cell types.

Further research on this compelling frontline will provide important context to the study of ceramides in the injured and diseased kidney. Nevertheless, the body of literature examining multiple models of kidney injury in vitro and in vivo suggests that ceramide accumulation is a common feature of AKI. Specifically, ceramides are implicated as drivers of mitochondrial dysfunction and cell death exacerbating kidney injury. Mechanistic studies are needed to delineate how ceramides are generated during AKI, which likely differ by injury etiology. Furthermore, interventions to critically test the benefit of ceramide-lowering for the prevention or treatment of AKI will provide exciting fodder for future therapeutic development.

Notably, the proposed mechanisms of ceramides implicated in AKI discussed in this review are likely relevant to chronic kidney disease.⁹⁹ We also acknowledge that alternative sphingolipid species have been implicated as mediators of kidney injury (eg, sphingosine-1-phosphate and glycosphingolipids). Discussion of these topics is outside the scope of the current review and has been reviewed exceptionally elsewhere.^9,100 We eagerly anticipate future developments in research involving ceramides and kidney injury to inform our understanding of the pathogenesis of AKI.