Mechanisms and Models of Kidney Tubular Necrosis and Nephron Loss

Abstract

Understanding nephron loss is a primary strategy for preventing CKD progression. Death of renal tubular cells may occur by apoptosis during developmental and regenerative processes. However, during AKI, the transition of AKI to CKD, sepsis-associated AKI, and kidney transplantation ferroptosis and necroptosis, two pathways associated with the loss of plasma membrane integrity, kill renal cells. This necrotic type of cell death is associated with an inflammatory response, which is referred to as necroinflammation. Importantly, the necroinflammatory response to cells that die by necroptosis may be fundamentally different from the tissue response to ferroptosis. Although mechanisms of ferroptosis and necroptosis have recently been investigated in detail, the cell death propagation during tubular necrosis, although described morphologically, remains incompletely understood. Here, we argue that a molecular switch downstream of tubular necrosis determines nephron regeneration versus nephron loss. Unraveling the details of this “switch” must include the inflammatory response to tubular necrosis and regenerative signals potentially controlled by inflammatory cells, including the stimulation of myofibroblasts as the origin of fibrosis. Understanding in detail the molecular switch and the inflammatory responses to tubular necrosis can inform the discussion of therapeutic options.

Kidney disease progression as defined by a decline in GFR is associated with tubular necrosis and nephron loss.^1–3 Derangements in podocytes and glomerular injury are involved in nephron loss according to some models,⁴ however, the massive changes in morphology require the death of tubular cells. Whereas apoptosis, the default pathway of the immunologically silent removal of cells, appears to be of limited relevance during nephron loss, the pathways of ferroptosis and necroptosis have recently attracted increasing attention. Our understanding is that no other cell death pathway explains tubular necrosis and the appearance of muddy brown casts in urine sediments better than ferroptosis.^5–12 The complete loss of a functional unit, such as an entire segment of the nephron, therefore, must be considered a possible mechanism underlying nephron loss. However, it is known from clinical observations that even significant tubular injuries can regenerate and recover. Inhibitors of necroptosis and ferroptosis are being investigated in clinical trials and might hold promise for the treatment of AKI, AKI to CKD transition, and the preservation of transplant kidney quality during transport on machine perfusion.

An Introduction to Regulated Cell Death

Apoptosis: Well Defined and Well Tolerated by the Immune System

Apoptosis is a complex form of cell death associated with exposure of high levels of phosphatidylserine (PtdSer) at the outer leaflet of the plasma membrane, serving as an eat-me signal for macrophages.¹³ Under nonapoptotic circumstances, distribution of PtdSer within the plasma membrane is regulated by flippases,¹⁴ such as ATP11A and ATP11C,¹⁵ that constantly keep PtdSer levels high in the inner leaflet.¹⁴ On caspase activation during apoptosis, the flippases are proteolytically processes and inactivated. At the same time, caspases activate the scramblase XKR8,¹⁶ rapidly increasing the PtdSer exposure on the cell surface where macrophages sense and engulf the apoptotic cells. Importantly, during apoptosis, the plasma membrane does not lose its integrity. Therefore, immunologic consequences of apoptosis are limited to the macrophage response and to the release of tissue messengers.¹⁷ Importantly, apoptotic cells do not release damage-associated molecular patterns (DAMPs), rendering apoptosis an immunologically silent death.^13,18 Apoptosis is therefore different from the other pathways described here, which are terminally executed by plasma membrane rupture, and therefore—in contrast with apoptosis—are necrotic in nature.

PtdSer is sensed by macrophages by PtdSer receptors such as the tissue injury molecule 1 (TIM-1),^19,20 also referred to as kidney injury molecule 1 (KIM-1).²¹ TIM-1 binding will start the process of engulfment of the bound membrane that may be an entire apoptotic cell, or a fragment of a necrotic cell that exposes high levels of PtdSer on the inner leaflet of the plasma membrane after the necrotic burst. In addition, KIM-1 was demonstrated to induce fatty-acid uptake in tubular epithelial cells.²² TIM-1/KIM-1 have been demonstrated to become highly upregulated and function as a biomarker during AKI²¹ associated with tubular necrosis and nephron loss.²³

It has been suggested that inhibition of apoptosis is protective in mouse models of AKI.²⁴ However, clear evidence for caspase activation in the tubular cell compartment outside the embryologic development is lacking, and caspase inhibitors have failed to protect mice from ischemia-reperfusion injury (IRI).²⁵ Along similar lines, deletion of the apoptosis-associated proteins caspase-8 or Fas-associated protein with death domain from renal tubules have not provided any protection from AKI.⁶ On the basis of these data and the plausibility of necrosis of renal tubules that induces KIM-1 expression, most authors agree the contribution of apoptosis to AKI in most animal models is neglectable (Box 1).

Box 1: Apoptosis in Sensu Strictu.

Outside the research field of cell death, the term “apoptosis” was commonly used as “cell death.” It has been referred to as a morphologic diagnosis by pathologists.²⁶ As originally introduced, however, apoptosis, like necrosis, reflects a morphologic pattern of fatal cell injury that has been expanded to include biochemical assays designed to improve its detection and quantification. Within this review, we refer to apoptosis as a caspase-dependent, noninflammatory cell death program with clearly defined features: apoptosis in sensu strictu.

Caspase-3 activation drives the apoptosis program rapidly²⁷ after proteolytic cleavage by caspase-8^28,29 (extrinsic apoptosis) or caspase-9^30–32 (intrinsic apoptosis). Cleaved caspase-3 is easily detectable in tissue slices by immunohistochemistry³³ or immunofluorescence.³⁴ In addition, cleaved caspase-3 can be detected by Western blotting and activity of caspase-3 by enzymatic assays.³⁵ Importantly, cleaved and active caspase-3 can also be found during forms of pyroptosis,³⁶ so the detection of cleaved caspase-3 is not sufficient to conclude on apoptosis. Likewise, the use of caspase-inhibitors (such as zVAD-fmk, qVD, or emricasan) will prevent apoptosis, but might also interfere with caspase-1 and/or caspase-11 during pyroptosis.

Another feature required for apoptosis is phospholipid flippase/scramblase activity. Anoctamin 6 (TMEM16F) is a phospholipid scramblase and a Ca-activated chloride channel involved in this process. Once PtdSer, which is expressed only on the inner leaflet of the plasma membrane in healthy cells, is exposed to the outer leaflet, it serves as an eat-me signal for macrophages that rapidly remove these cells in an immunologically silent manner.¹³

Another process that is inevitably associated with apoptosis is the shrinking of the cells, followed by the process of membrane blebbing, best detected in time-lapse imaging. However, ferroptotic cells may also show features of membrane blebbing, which exemplifies the need for independent assay to conclude on apoptosis.¹³

Apoptosis is further characterized by nuclear chromatin condensation and a phenomenon referred to as DNA laddering. It should not be concluded on apoptosis if the features mentioned here have not been fully investigated.

Finally, all known regulated cell death pathways result in the positivity of terminal deoxynucleotidyl transferase dUTP nick end labeling staining, which simply detects double-strand breaks in the DNA, and therefore is nonspecific.

Ferroptosis: Iron-dependent Necrosis by Lipid Peroxidation

A growing list of manuscripts suggest ferroptosis critically contributes to AKI (Table 1). Ferroptosis is initiated by the failure to control lipid peroxidation. In physiologic circumstances, several molecular surveilling systems prevent ferroptosis. The best studied system centrally depends on the function of the glutathione peroxidase 4 (GPX4), a glutathione (GSH) metabolizing selenoenzyme that turns oxidized lipids to respective inactivated alcohols (Figure 1).^37,38 A second layer of regulation for this system is the intracellular concentration of GSH, which is provided in different ways in a cell type–specific manner, for example, by the transmembrane glutamate/cystine antiporter system Xc-, the target of the type 1 ferroptosis inducer erastin, in combination with the GSH synthase.^39–41 Other pathways that regulate the GSH pool include the trans-sulfuration pathway⁴² and the dipeptidase-1.⁴³

Table 1.

Evidence for the role of ferroptosis in AKI

Number	Date of publication	Authors	Journal	Title	Type of Evidence— AKI Model	Reference— PMID
1	2014, November	Linkermann et al.	PNAS	Synchronized renal tubular cell death involves ferroptosis	Pharmacologic (IRI)	25385600
2	2014, December	Friedman-Angeli et al.	Nature Cell Biology	Inactivation of the ferroptosis regulator Gpx4 triggers acute renal failure in mice	Genetic (spontaneous)	25402683
3	2017, January	Martin-Sanchez et al.	JASN	Ferroptosis, but not necroptosis, is important in nephrotoxic folic acid-induced AKI	Pharmacologic (FA-AKI)	27352622
4	2017, June	Martens et al.	Cell Death and Disease	Sorafenib tosylate inhibits directly necrosome complex formation and protects in mouse models of inflammation and tissue injury	Pharmacologic (IRI)	28661484
5	2017, October	Müller et al.	Cell Mol Life Sci	Necroptosis and ferroptosis are alternative cell death pathways that operate in acute kidney failure	Histologic (ACSL4-staining) in IRI	28551825
6	2018, May	Adedoyin et al.	AJP-renal	Heme oxygenase-1 mitigates ferroptosis in renal proximal tubule cells	Pharmacologic (PTCs)	28515173
7	2018, May	Stoppe et al.	Science Translational Medicine	The protective role of macrophage migration inhibitory factor in acute kidney injury after cardiac surgery	Pharmacologic (rhabdomyolysis-induced AKI)	29769287
8	2019, April	Guerrero-Hue et al.	FASEB Journal	Curcumin reduces renal damage associated with rhabdomyolysis by decreasing ferroptosis-mediated cell death	Pharmacologic (rhabdomyolysis-induced AKI)	31034781
9	2019, June	Mishima et al.	JASN	Drugs repurposed as antiferroptosis agents suppress organ damage, including AKI, by functioning as lipid peroxyl radical scavengers	Pharmacologic (CP-AKI)	31767624
10	2019, June	Huang et al.	J Cell Mol Med	Augmenter of liver regeneration protects the kidney from ischemia-reperfusion injury in ferroptosis	Pharmacologic (IRI)	30993878
11	2019, November	Deng et al.	JCI	Myo-inositol oxygenase expression profile modulates pathogenic ferroptosis in the renal proximal tubule	Pharmacologic (CP-AKI)	31437128
12	2019, December	Su et al.	JBC	Pannexin 1 mediates ferroptosis that contributes to renal ischemia/reperfusion injury	IRI	31694915
13	2020, January	Hu et al.	Cell Death and Disease	VDR activation attenuate cisplatin induced AKI by inhibiting ferroptosis	Pharmacologic (CP-AKI)	31996668
14	2020, February	Córdoba-David et al.	Scientific Reports	Effective nephroprotection against acute kidney injury with a star-shaped polyglutamate-curcuminoid conjugate	Pharmacologic (tubular cells)	32029842
15	2020, February	Lee et al.	Nature Cell Biology	Energy-stress-mediated AMPK activation inhibits ferroptosis	IRI	32029897
16	2020, July	Wang et al.	Journal of Advanced Research	Quercetin alleviates acute kidney injury by inhibiting ferroptosis	Pharmacologic (IRI, FI-AKI)	33364059
17	2020, August	Zhao et al.	Cell Death and Disease	XJB-5–131 inhibited ferroptosis in tubular epithelial cells after ischemia-reperfusion injury	Pharmacologic (IRI, FI-AKI)	32796819
18	2020, October	Ding et al.	Cell Death and Disease	miR-182–5p and miR-378a-3p regulate ferroptosis in I/R-induced renal injury	Pharmacologic (IRI)	33116120
19	2020, November	Li et al.	Free Radic Biol Med	Inhibition of ferroptosis by up-regulating Nrf2 delayed the progression of diabetic nephropathy	Pharmacologic (diabetic nephropathy)	33152439
20	2020, November	Yant et al.	Life Sci	Tocilizumab mimotope alleviates kidney injury and fibrosis by inhibiting IL-6 signaling and ferroptosis in UUO model	Pharmacologic (UUO)	32979361
21	2020, December	Wang et al.	Eur J Pharmacology	Ferroptosis involves in renal tubular cell death in diabetic nephropathy	Pharmacologic (diabetic nephropathy)	32976829
22	2021, January	Chen et al.	Cell Death and Disease	Legumain promotes tubular ferroptosis by facilitating chaperone-mediated autophagy of GPX4 in AKI	Pharmacologic (IRI, FA-AKI)	33431801
23	2021, January	Guo et al.	Br J Pharmacology	Targeted inhibition of Rev-erb-α/β limits ferroptosis to ameliorate folic acid-induced acute kidney injury	Pharmacologic (FA-AKI)	33068011
24	2021, January	Jiang et al.	Mol Med Rep	Effects and molecular mechanism of pachymic acid on ferroptosis in renal ischemia reperfusion injury	Pharmacologic (IRI)	33215224
25	2021, January	Deng et al.	Front Pharmacol	Mitochondrial iron overload-mediated inhibition of Nrf2-HO-1/GPX4 assisted ALI-induced nephrotoxicity	Pharmacologic (AAN)	33584308
26	2021, February	Zhang et al.	J Cell Physiol	Involvement of GPX4 in irisin's protection against ischemia reperfusion-induced acute kidney injury	Pharmacologic (IRI)	32583428
27	2021, February	Ma et al.	Dig Dis Sci	Inhibition of ferroptosis attenuates acute kidney injury in rats with severe acute pancreatitis	Pharmacologic (pancreatitis-induced AKI)	32219613
28	2021, February	Kim et al.	Cell Death and Disease	Characterization of ferroptosis in kidney tubular cell death under diabetic conditions	Pharmacologic (diabetic nephropathy)	33558472
29	2021, February	Feng et al.	Front Endocrinol	Ferroptosis enhanced diabetic renal tubular injury via HIF-1α/HO-1 pathway in db/db mice	Pharmacologic (diabetic nephropathy)	33679620
30	2021, March	Li et al.	Br J Pharmacology	Nuciferine protects against folic acid-induced acute kidney injury by inhibiting ferroptosis	Pharmacologic (FA-AKI)	33450067
31	2021, April	Yang et al.	Clin Transl Med	Dimethyl fumarate prevents ferroptosis to attenuate acute kidney injury by acting on NRF2	Pharmacologic (FA-AKI, CP-AKI)	33931960
32	2021, July	Tonnus et al.	Nature Communications	Dysfunction of the key ferroptosis-surveilling systems hypersensitizes mice to massive tubular necrosis during acute kidney injury	Genetic (GPX4-inactivation, FSP1-ko) and Pharmacologic	34285234
33	2021, September	Zhao et al.	Free Radic Biol Med	Iron deficiency exacerbates cisplatin- or rhabdomyolysis-induced acute kidney injury through promoting iron-catalyzed oxidative damage	Pharmacologic (CP-AKI)—patient data	34298093
34	2021, November	Zhao et al.	Free Radic Biol Med	Endoplasmic reticulum stress-mediated autophagy activation is involved in cadmium-induced ferroptosis of renal tubular epithelial cells	Pharmacologic (cadmium-induced AKI)	34520822
35	2021, September	Lin et al.	Biomedicines	Nephroprotective role of chrysophanol in hypoxia/reoxygenation-induced renal cell damage via apoptosis, ER stress, and ferroptosis	Hypoxia/reoxygenation of tubular epithelial cells	34572468

FA, folic acid; PTC, proximal tubular cells; CP, cisplatin; VDR, Vitamin D receptor; FA, foliac acid; UUO, unilateral ureteral obstruction; AAN, aristolochic acid nephropathy; IRI, ischemia-reperfusion injury.

Figure 1.

Pathways of renal cell death. Ferroptosis (left) is a failsafe rather than a typical cell death pathway. In cellular homeostasis, H₂O₂ concentrations and iron-catalyzed Fenton reactions are limited by diverse cellular antiredox systems. The best studied system relies on GSH, which is generated intracellularly and depends on supply via system Xc-, a cys/glu antiporter in the plasma membrane, or products of the trans-sulfuration pathway. On sufficient GSH concentrations, GPX4 prevents lipid peroxidation that otherwise leads to plasma membrane rupture by unknown mechanisms. In contrast, the oxidoreductase FSP1 (also known as AIFM2) prevents lipid peroxidation on myristoylation-dependent recruitment to the plasma membrane in a GSH-independent manner. Ferroptosis occurs as a noncell-autonomous pathway in a process referred to as synchronized regulated necrosis (SRN). Apoptosis (center) represents a noninflammatory pathway that is mediated by caspases. Two distinct signaling pathways of apoptosis, extrinsic and intrinsic apoptosis, have been characterized. In extrinsic apoptosis, death receptors such as TNFR1, CD95 (Fas), and TRAIL-R, through the engagement of various intracellular adaptor proteins (Fas-associated protein with death domain, TRADD) lead to the activation of the master regulator caspase-8. On homodimerization, caspase-8 cleaves the effector caspases-3/-6/-7 to propagate the apoptosis program. On loss of mitochondrial outer membrane potential (MOMP) and the BAX-BAK–mediated release of cytochrome c from the mitochondria into the cytosol, the intrinsic apoptotic pathway is triggered. Within the cytosol, cytochrome c, APAF1 and caspase-9 form the apoptosome, which activates the effector caspases-3/-6/-7. The typical morphology includes nuclear condensation, early loss of cellular volume (shrinking), subsequent membrane blebbing, and exposure of PtdSer. Importantly, PtdSer exposure represents an eat-me signal to macrophages that eliminate apoptotic cells. Importantly, the plasma membrane does not lose its integrity during this process. Whereas apoptosis depends on the activation of caspases, necroptosis (right) is mediated by kinases. Depending on its RHIM domain, RIPK3 forms an amyloid-like structure referred to as the necrosome, the central signaling platform of necroptosis. Therein, RIPK3 phosphorylates the pseudokinase MLKL. By unknown mechanisms, pMLKL triggers a plasma membrane rupture, a process that was demonstrated to be counteracted by the membrane repair ESCRT-III complex. The necrosome can be engaged by death receptor signaling in the condition in which caspases are absent or inhibited (e.g., by viral proteins), and RIPK1 no longer intercalates with RIPK3. Other ways of engaging the necrosome are through TLRs via the RHIM-containing adaptor molecule TRIF, or by activation of the protein ZBP1/DAI in response to sensing intracellular oligonucleotides.

A second GSH-independent system centrally involves the ferroptosis suppressor protein 1 (FSP1).^44–47 As GPX4, this oxidoreductase system is also important in the kidney because FSP1-deficient mice are sensitive to IRI.¹² Alongside GPX4 and FSP1, the regulation of the intracellular iron pool, for example, by hepcidin⁴⁸ or the heme oxidase 1 and H-ferritin,^49,50 the machinery involved in H₂O₂ control (e.g., p450 oxidoreductase⁵¹), and lipid synthesis enzymes such as ACSL4^52–54 are critically involved in ferroptosis regulation. Considering the importance of ferroptosis and the rapid movement of the field,^55–60 it should be stated the kidney was the first organ shown to undergo ferroptosis, and has evolved as a model system for ferroptosis researchers. However, the detailed emerging mechanism of ferroptosis is beyond the scope of this review.^46,61–69 It is important to understand our novel concept of the limited immunogenicity of ferroptosis to consider that in contrast to apoptosis, necroptosis, and pyroptosis, ferroptotic cells release high amounts of oxidized phospholipids, such as oxidized phosphatidylethanolamine, oxidized PtdSer, and oxidized phosphatidylinositol.⁷⁰

Necroptosis: Mixed-lineage Kinase Domain Like-mediated Necrosis To Defend Viruses

Necroptosis (Figure 1) is mediated by RIPK3-dependent phosphorylation^71–73 of the pseudokinase mixed-lineage kinase domain like (MLKL) and subsequent plasma membrane rupture.^74,75 RIPK3 can be activated by at least three systems. First, death receptors of the TNF superfamily, such as TNFR1, form a TNF-receptor signaling complex to unleash the pronecroptotic function of the kinase RIPK1.^76,77 Human mutations in RIPK1 have been described to lead to an autoinflammatory disease.⁷⁸ The only known target of RIPK1 is RIPK1 itself, but the phosphorylation at the p166 residue regulates an autoinhibitory function^77,79–82 that prevents RIPK3 oligomerization via an RIP homotypic interacting motif (RHIM)-domain in both RIPK1 and RIPK3.^83–87 Phosphorylation of RIPK1, therefore, drives the necroptosis machinery. Second, all toll-like receptors that engage the intracellular adapter protein TRIF can stimulate RIPK3 oligomerization and phosphorylation in an RIPK1-independent manner.^88–92 Finally, intracellular nucleotides (such as misfolded zRNA) sensed by the protein ZBP1 results in RIPK3-dependent necroptosis as well.^93–96 It is important for the purpose of this review that necroptosis potently simulates dendritic cell (DC)–dependent crosspriming of effector T cells.⁹⁷ In addition, necroptosis is associated with the maturation and release of cytokines.^78,98

The necroptosis system is of importance for AKI,^99–101 kidney transplantation,^102–104 and potentially for nephron loss (see below). MLKL is among the most significantly upregulated genes in a number of models of AKI,¹⁰⁵ and RIPK3-¹⁰⁰ and MLKL¹⁰¹-deficient mice are protected from IRI in the clamp ischemia model. In human ANCA-vasculitis samples¹⁰⁶ and kidney transplant biopsies,¹⁰¹ individual tubular cells stain positive for a highly specific antibody against phosphorylated MLKL (pMLKL). In addition, pMLKL positivity was recognized particularly in peritubular capillaries after transplantation.¹⁰⁷ Future studies will investigate if the expression of RIPK3 and/or MLKL in kidney biopsies is helpful for risk stratification. As for the ferrostatins, small molecule inhibitors of necroptosis (necrostatins) are readily available for clinical trials.

DAMPs: Cell-death–driven Consequences for the Immune System

Necroinflammation is a process in which necrotic cells release DAMPs and thereby shape the immune response to the necrotic debris.^108–110 Exposing necroptotic cell antigens to conventional DCs (cDC1) stimulates crosspresentation to CD8-positive cytotoxic T cells.¹¹¹ Crosspresentation of necrotic antigens, such as actin,¹¹² is regulated in a complex manner. In the case of actin, the cDC1 receptor DNGR1 is required for crosspresentation.^113,114 At least in the microenvironment of cancers, however, soluble factors, such as gelsolin, are known to interfere with actin binding to DNGR1 to prevent antitumor immunity.¹¹¹ Similar mechanisms are likely to contribute to necroinflammation in the kidney, but have not been investigated so far. In summary, the necrotic death of kidney cells not only results a loss of function, but leaves the remaining tissue with potentially immunogenic necrotic debris. We are only just beginning to understand how distinct necrotic cell death pathways shape the immune system in different ways.^57,97 There is, however, reason to believe cDC-mediated crosspriming of T cells after necroptosis is associated with significant DC-to-myofibroblast crosstalk and potentially myofibroblast proliferation.¹¹⁵ Taken together, as depicted in Figure 2, necroptosis drives the adaptive immune system, which explains the proliferation of fibroblasts to cause fibrosis and irreversible nephron loss (see below).

Figure 2.

A hypothetical model of the interplay between the immune system and regulated cell death pathways. Although under fundamental debate, ferroptosis may create an anti-inflammatory environment, at least for the adaptive immune system. Although T cell activation, T cell crosspriming, and B cell activation may be inhibited by ferroptosis, cells of the innate immune system such as macrophages and neutrophils may better resist the environmental scenario of high lipid peroxidation and redox imbalance to remove necrotic debris, a feature that potentially allows necrotic kidney tubules to regenerate. In such a model, profibrotic conditions would be antagonized by the innate immune system through the removal of necrotic debris. In contrast with all other pathways of regulated cell death, classic apoptosis does not result in plasma membrane rupture and therefore does not expose intracellular epitopes to the immune system. PtdSer exposure on the surface of apoptotic cells functions as an eat-me signal and the immunologically silent removal of PtdSer-exposing cells. Necroptosis is interpreted as a defense mechanism against virus-carrying cells. During necroptosis (and most likely pyroptosis), the initiation of DC activation and T cell crosspriming alongside with the maturation of cytokines establishes a solid adaptive immune response and a vaccination against viral epitopes. Maladaptive repair and the activation of myofibroblasts in the kidney are likely a consequence of adaptive immune cell activation and nonferroptotic, nonapoptotic regulated cell death. We clearly point out that the presented model is speculative and serves as a working model that needs to be tested in the future.

In contrast with necroptotic death of renal tubular epithelial cells, ferroptosis may not trigger the adaptive immune response despite the release of DAMPs. In contrast with previous assumptions that were on the basis of the DAMP release as the exclusive immune-regulatory factor in ferroptosis,^116–118 including our own previously published model,¹⁰⁹ convincing accumulating data suggest anti-immunogenic factors are key in ferroptosis. Indeed, it was recently demonstrated that T cells are paralyzed by myeloid cells in high reactive oxygen species conditions.¹¹⁹ This important effect is mechanistically mediated as a direct cell-to-cell contact between myeloid cells and T cells.¹¹⁹ This mechanism may explain why, despite massive amounts of tubular necrosis, hardly any adaptive inflammatory response is seen in histologic samples of IRI-induced acute tubular necrosis.^12,101 It is important to consider that in this scenario, the innate immune system and myeloid-derived suppressor cells are active, explaining the infiltration of macrophages and neutrophils (Figure 2). It is a new hypothesis that a microenvironment in which adaptive immunity is actively blocked may allow tubular regeneration. Of course, this model is hypothetical and future work should allow the testing of this hypothesis. However, these considerations are in line with the absence of an adaptive immune cell infiltration in GPX4-dysfunctional mice.⁵

Kidney Tubular Necrosis: Cell Death Propagation in the Kidney

Primary Tubular Injury

Necroptosis and ferroptosis have been convincingly demonstrated by different laboratories to contribute to tubular necrosis (see above). In most models of AKI, however, the relative contribution of these regulated necrotic cell death pathways clearly favors a major role for ferroptosis.^120,121 Indeed, entire segments of proximal tubules and thick ascending limbs undergo ferroptosis.^6,12 During this process, lipid peroxidation dominates the injured microenvironment between the basal laminas of the tubules and may paralyze the adaptive immune system, thereby preventing the proliferation of myofibroblasts.¹¹⁹

In isolated renal tubules, spontaneous release of lactate dehydrogenase was demonstrated two decades ago.^122–124 A significant portion of this spontaneous lactate dehydrogenase release is accounted for by ferroptosis.¹² As briefly mentioned above, ferroptosis is a cell death modality that typically involves noncell autonomous cell death propagation.^5,6,125,126 Similar effects have been observed in the fins of zebrafish.^127,128 This novel and expanding of field noncell autonomous cell death propagation is testing several hypotheses, some of which are depicted in Figure 3. One important aspect is the shared cytosol between connected tubular compartments, especially in its redox capacity. NADP(H) concentrations may vary along the renal tubule and become insufficient in preventing lipid peroxidation, a hallmark of ferroptosis.¹²⁹ In addition, an accompanying wave of calcium was described in the cell culture, and might contribute to the cell death propagation in the renal tubules. Such factors are included in Figure 3 and account for regulatory mechanisms of the pace of cell death propagation. Future work will be required to understand these intercellular signals in more detail, to identify potential novel therapeutic targets.

Figure 3.

SRN: cell-death propagation during renal tubule ferroptosis. In most cases, it is unclear how a necrotic zone of dead cells propagates during an event of sepsis or ischemia. On the basis of time-lapse videos obtained from isolated perfused renal tubules that underwent ferroptosis, the concept of SRN emerged.⁶ Although this phenomenon has not been observed in humans, necrotic tubules in the urine of patients and experimental intravital microscopy videos suggest cell death occurs in a “wave of death.”^6,126,128 Given the connections between cells in a functional syncytium such as a renal tubule, the adrenal gland, or the myocardium, it is tempting to speculate that the intracellular redox capacity (NADPH concentration) diffuses through intercellular junctions. If the redox capacity decreases in a dying cell, an NADPH gradient forms, which renders the closest neighbor at high risk of undergoing ferroptosis. It is now clear the bulk necrotic area that occurs during myocardial infarction or tubular necrosis originates from cells that underwent ferroptosis. We speculate that similar mechanisms might occur in other organs, such as the brain upon stroke, or the adrenal glands during sepsis or shock.

In contrast with ferroptosis, death of a tubular cell by necroptosis or apoptosis is much less common. Apoptosis of tubular cells was convincingly described during embryonic development,¹³⁰ but genetically deleting key apoptosis proteins, such as caspase-8 or Fas-associated protein with death domain, in adult mice does not affect tubular necrosis patterns in AKI models or isolated kidney tubules.⁶ As described for several cancers, apoptotic programs in tubular cells may be reactivated during tubular regeneration. This would explain the detection of cleaved-caspase 3 signals in immunofluorescence or immunohistochemistry in some models of AKI. However, despite intensive unpublished investigations, in our own laboratory we failed to detect a single cleaved caspase-3 positive tubular cell in adult mice or human tissue in two decades of cell death research. In contrast, in addition to ferroptosis, pMLKL signals in renal tubular cells indeed indicate necroptosis, and understanding the appearance of those particularly immunogenic cells will be key to unraveling the mechanisms of nephron loss.

Tubular Necrosis Secondary to Peritubular Capillary Dysfunction

In contrast with tubular cells that predominantly die by ferroptosis, peritubular capillaries appear to follow a very different set of cell death modalities. The apoptosis/necroptosis system is known to occur in endothelial cells¹³¹ and peritubular capillaries,¹⁰⁴ and drive capillary rarefaction.¹⁰⁴ In those IRI studies, a nonspecific RIPK3 signal by immunohistochemistry still demonstrated tubules die in a typical synchronized manner, indicating secondary ferroptosis after peritubular capillary dysfunction.¹⁰⁴ More experiments directed to cell death of peritubular capillaries are required to shed light on the detailed mechanism of secondary tubular necrosis. However, such experiments are tremendously important for understanding nephron loss. Some recent evidence from single-cell RNA sequencing analyses has been published and may provide the first insights into endothelial regulation.^132,133 Such approaches, despite being entirely descriptive in nature and mechanistically insufficient to make conclusions about the cell death pathways involved, may indicate which mRNAs associated with transcription to the relevant cell death proteins are involved. Importantly, again, ferroptosis appears as a key mechanism in one of those studies.¹³²

Nephron Loss

A Model of Factors that Regulate Nephron Loss

Nephron loss has been broadly discussed over the past few decades.¹³⁴ However, only in selected animal models, such as of GMB nephritis, necroptotic cell death of neutrophils and the release of neutrophil extracellular traps cause the glomerular filter to collapse.¹⁰⁶ Other models discuss protein leakage into the primary urine, but how excessive reabsorption of proteins may cause tubular necrosis remains obscure, although reactive oxygen species may be involved. On an extensive literature search, “abnormal filtrate spreading” or “obstruction by overgrowth” fail to provide clear mechanistic insights into how tubular cells could die.¹³⁴ More generally, the concept of the loss of nephrons as a consequence of glomerular injury fails to explain AKI to CKD progression,³ as recently reviewed.¹³⁵ The oversimplified view in which a tubule dies on glomerular dysfunction, therefore, appears unsustainable. In contrast, genetic destruction of renal tubules, for example, by deletion of GPX4, results in nephron loss by ferroptosis⁵ without the glomeruli being affected in any way. It is, however, entirely unclear how necrotic casts would form after tubular necrosis and how changes in tubular flow affect this process.

All data discussed above suggest acute tubular necrosis is reversible in some patients, but not in others. Therefore, the molecular switch, in other words, the decision of recovery or maladaptive repair and subsequent fibrosis, is downstream of tubular necrosis. Although to the best of our knowledge, nothing has been convincingly demonstrated on the nature of this decision process, in this study we propose a novel working hypothesis. On the basis of the effects on crosspriming and potentially the stimulation of myofibroblast proliferation, necroptosis may be upstream of maladaptive repair, fibrosis, and the clinical picture of interstitial fibrosis and tubular atrophy. In contrast, ferroptosis-mediated inhibition of the adaptive immune system might allow for recovery of the injured tissue. This hypothesis, however, cannot be interpreted as more than a working model, but we consider it conclusive that the cell death signal and the respective necroinflammatory response contribute to the divergent outcome. Considering these thoughts on a more general scale, ferroptosis may represent a means to allow regeneration of renal tubules. In models of hypoperfusion, such as IRI or septic AKI, tubular ferroptosis dominates. Similarly, a growing body of literature supports that toxic AKI is mediated by ferroptosis. In conclusion, in most cases, nephron loss is more likely to be explained by primarily tubular injury, or as a result of hypoperfusion in peritubular capillaries, secondary tubular injury. Given the insufficient quality of preclinical models of AKI-induced CKD progression, reliable data on the basis of animal models of nephron loss are sparse. Until today, no intravital microscopy-based detection of acute tubular necrosis of more than one cell¹³⁶ has been reported. Along similar lines, nephron loss is a morphologic explanation for AKI to CKD progression. The number of lost nephrons may correlate with GFR decline. Even the widely accepted thoughts on age-related nephron loss methodologically relies on autopsy kidneys or enhanced computed tomography and renal biopsy analysis.^137,138 The latter method may be of clinical importance to identify individuals at risk for CKD progression on the basis of low nephron endowment.¹³⁹

Specific Considerations regarding Nephron Loss in Kidney Transplants

Beyond general regulators of nephron loss, kidney allografts are at risk for nephron loss as a result of additional specific factors. These include HLA incompatibility–driven inflammation, such as in acute and chronic antibody- or T cell–mediated rejection, but also nonimmunologic components. Allograft artery stenosis or thrombosis and peritransplant injuries, recipient arterial hypertension, and hyperlipidemia and general cardiovascular risk factors commonly contribute to hypoperfusion and to an insufficient partial oxygen pressure. Specific necrotic signals may result from toxicities, for example, side effects of calcineurin inhibitors (CNIs), or from the defense against virally infected tubular cells. CNIs have been demonstrated to cause characteristic tubular morphologic changes in response to necroptosis-regulating signaling, for example, through FN14, a TNFR-superfamily member.¹⁴⁰ In contrast with CNI-induced tubular damage, cytomegalovirus infections are known to be cleared by necroptotic signaling.^141,142 Along similar lines, BK virus infections are known to be accompanied by necrotic tubular damage^143–145 that might be caused by necroptosis. It is possible that necroptosis of virally infected cells, according to the model presented in Figure 2, drives myofibroblast proliferation and fibrosis. Finally, after a long cold ischemia time before anastomosis, it is possible nephron loss occurs within the first hours of the reperfusion phase. In such a scenario, addition of necrostatins and/or ferrostatins to a kidney transplant machine perfusate may prevent nephron loss and associated necroinflammation.

Therapeutic Strategies

Food and Drug Administration–approved Drugs that target Ferroptosis and Necroptosis

Several Food and Drug Administration–approved drugs function as cell death inhibitors. In the case of ferroptosis, Mishima et al. published a comprehensive analysis on promethazine, omeprazole, carvedilol, estradiol, rifampicin, indole-3-carbinol, propranolol, and thyroid hormones to function as peroxyl radical scavengers.¹⁰ Although the antiferroptotic potency of these drugs is much lower compared with modern ferroptosis inhibitors, such as 3203¹⁴⁶ or liproxstatin-1,¹⁴⁷ it cannot be excluded that relevant antiferroptotic effects result from their application. Interestingly, IL-4–induced-1 (IL-4i1), an amino acid oxidase secreted from immune cells with high homology to snake venoms (L-amino acid oxidases), exerts antiferroptotic functions by generation of indole-3-pyruvate from tryptophan.¹⁴⁸ These novel mechanisms have not been therapeutically targeted until today.

As with repurposed drugs that inhibit ferroptosis, the anticonvulsant phenytoin functions as an antinecroptotic compound,¹⁰¹ although much less potent compared with specific RIPK1 inhibitors. Along similar lines, the anticancer therapeutics ponatinib^149,150and pazopanib¹⁴⁹ were also demonstrated to function as inhibitors of RIP kinases.¹⁵¹ Most intriguingly, however, the original necrostatin (necrostatin-1, Nec-1^152,153) contains a thiohydantoin motif base that inhibits ferroptosis by approximately 20 µM.⁵ Therefore, it is possible to generate combined small molecule inhibitors of necroptosis and ferroptosis as recently demonstrated by the generation of Nec-1f (where “f” means inhibition of ferroptosis).¹²

Potential Side Effects of Cell Death Inhibitors

It may not matter to a single cell how it dies, but it affects the surrounding tissue and the immune system. Inhibiting one cell death pathway, such as apoptosis, may allow a more immunogenic cell death modality to pursue, for example, necroptosis.^154,155 This has been most clearly demonstrated for the proapoptotic caspase-8, which cleaves caspase-3 to execute apoptosis and at the same time, through a caspase cleavage site in RIPK1, inhibits necroptosis.⁸⁵ Although similar connections between regulated cell death pathways have been demonstrated for an apoptosis-to-pyroptosis crosstalk,^85,156–158 nothing is known in respect to the interconnectivity of any cell death pathway to ferroptosis. Importantly, however, crossregulation of ferroptosis and other pathways cannot be excluded. In conclusion, therapeutically interfering with the web of cell death pathways may favor other cell death pathways that could potentially be even more immunogenic. In cancer immunotherapy, this might more effectively treat the cancer, yet would be potentially dangerous in the case of solid organ transplantation.

As mentioned above, necroptosis is a means to defend viruses, and cytomegalovirus has already found a way to protect itself not only against apoptosis, but also against necroptosis through the expression of the viral protein M45.^141,159,160 Other viruses may copy such strategies. That said, simultaneous inhibition of the critical pathways may be required for the best outcomes and potentially disease-specific and individualized cell death therapies. We consider it likely that solid organ transplant ex-vivo perfusion systems will be safe to test ferrostatins for clinical purposes.

The model on nephron loss (Figure 2) is speculative. However, as all scientists should, we worked out a hypothesis using the best of our knowledge. Every nephrologist who once saw the massive T cell infiltration into a graft on acute T cell–mediated rejection realizes the inflammatory capacity of our immune system. Despite massive tubular necrosis, nothing comparable happens to kidneys that underwent IRI. This simple consideration, to us, rules out that ferroptosis is a highly immunogenic cell death pathway. Future work should aim at investigating this hypothesis, for the sake of progression of knowledge on nephron loss.

Disclosures

A. Linkermann issued a patent for Nec-1f, a combined inhibitor of necroptosis and ferroptosis (#20160943.5); reports having consultancy agreements with Alexion, Genentech, and HBM; reports receiving research funding from Apogenix, Fresenius, Novartis, and Pfizer; reports receiving honoraria from Alexion, Genentech, and Novartis; reports being a scientific advisor or member of AJP renal, the American Journal of Transplantation, Cell Death and Differentiation Guest Editor, Cell Death and Disease, Cellular and Molecular Life Sciences Guest Editor, Clinical Kidney Journal, JASN, Molecular and Cellular Oncology, Oncotarget, the Russian Journal of Internal Medicine, and Seminars in Nephrology Guest Editor. All remaining authors have nothing to disclose. We acknowledge funding to our laboratory by the Medical Clinic 3, University Hospital Carl Gustav Carus Dresden, Germany, by the German research foundation (DFG) SFB-TRR 205, SFB-TRR 127, the international research training group 2251, the priority program on ferroptosis (SPP 3206) and by the Wilhelm-Sander-Foundation (2020.043.1)

Funding

This work was supported by the Heisenberg Professorship to A. Linkermann (project number 324141047).