INNATE IMMUNITY IN NEPHROTOXIC ACUTE KIDNEY INJURY

Abstract

Acute kidney injury (AKI) is common among hospitalized patients and is associated with high morbidity and mortality. Inflammation is recognized to play an important role in both ischemic and toxic models of AKI. Cisplatin is a widely used and highly effective cancer chemotherapeutic agent but carries the risk of nephrotoxicity. We have used a model of cisplatin-induced AKI to explore the functions of the innate immune response in kidney injury. Several components of innate immunity, such as Toll-like receptor sensing and inflammatory cytokine production, contribute to both ischemic and cisplatin-induced AKI. Importantly, it is the activity of these components in kidney parenchymal cells, rather than immune cells, which mediate AKI. Cellular components of innate immunity, such as neutrophils and dendritic cells, appear to play disparate roles in ischemic vs toxic AKI. Innate immune pathways could be targeted to prevent or treat AKI.

INTRODUCTION

Acute kidney injury (AKI) occurs in 5% to 7% of hospitalized patients (1,2) and results in a mortality rate of approximately 50% (3). The financial costs of AKI are estimated to be 8 billion dollars per year, or approximately $130,000 per life-year saved (4). In addition to its short-term consequences, AKI also increases the risk and rate of progression for chronic kidney disease and perhaps cardiovascular events (5). Nephrotoxic drugs are a common cause of AKI. One example of nephrotoxin AKI is cisplatin nephrotoxicity (6). Cisplatin (cis-diamminedichloroplatinum[II]) is an antineoplastic drug introduced into clinical use in the US 40 years ago. The main dose-limiting side effect of cisplatin is nephrotoxicity (7-9). Yet, despite intense efforts over the ensuing decades to find less toxic but equally effective alternatives, cisplatin continues to be a standard component of treatment regimens for a variety of tumors, including those found in the head and neck, testicle, lung, ovary, and breast (6). AKI occurs in approximately 10% to 15% of patients during their first encounter with the drug (10) and in about one-third of all patients at some point during their course. The clinical features which tend to distinguish cisplatin-induced AKI from other forms of AKI include polyuria, magnesium wasting, potassium wasting, and occasionally Fanconi syndrome. Although we think of cisplatin-induced AKI as an acute process, recovery may be incomplete, particularly after repeated courses, and result in long-term loss of renal function (11).

MECHANISMS OF CISPLATIN-INDUCED AKI

Cisplatin is cleared by the kidney by both glomerular filtration and tubular secretion (12). Cisplatin concentrations within the kidney exceed those in blood suggesting an active accumulation of drug by renal parenchymal cells which may contribute to the susceptibility of the kidney to cisplatin injury. Two transporters, organic cation transporter 2 (13) and a copper transporter (14), may mediate cisplatin uptake into proximal tubule cells. Once in the kidney, cisplatin is thought to produce cell injury through multiple mechanisms including production of reactive oxygen species, damage to nuclear and mitochondrial DNA, activation of mitogen-activated protein kinases and activation of both intrinsic and extrinsic apoptotic pathways (6,15). In addition to direct cellular toxicity, there is growing recognition of the importance of inflammation in the pathogenesis of cisplatin- induced AKI.

INNATE IMMUNITY IN CISPLATIN-INDUCED AKI

The innate immune system provides for the rapid but nonselective response to invading organisms or tissue injury, and accounts for the classic features of inflammation. Mediators of the innate immune response include sentinel cells which sense danger, effector cells, pattern recognition molecules, and a variety of soluble mediators, such as cytokines, chemokines, complement, and prostaglandins. We have used the single-dose model of cisplatin toxicity popularized by Arany and Safirstein (16) to explore the role of innate immunity in cisplatin nephrotoxicity. In this model, a single dose of cisplatin (20 mg/kg) produces severe kidney injury 48 to 72 hours after injection in normal mice.

TUMOR NECROSIS FACTOR

Tumor necrosis factor (TNF) is a potent proinflammatory cytokine that plays important roles in chronic inflammation and autoimmune diseases such as rheumatoid arthritis, autoimmune diabetes, and multiple sclerosis (2,10). In earlier work we have shown that both serum and kidney levels of TNF were increased after cisplatin treatment in mice (17). This increase was due to both increased stability of TNF mRNA in the kidney (18) and increased translation through a p38 mitogen-activated protein kinase–dependent pathway (19). To address the functional relevance of TNF in the pathogenesis of cisplatin-induced acute renal failure, renal function and renal histology were examined in mice treated with cisplatin in the presence or absence of TNF inhibitors and also in TNF knockout mice (17). Treatment with TNF inhibitors or genetic deletion of TNF reduced cisplatin-induced renal dysfunction and also reduced histologic evidence of injury (17). These results, which have been confirmed by a number of laboratories (20,21), establish an important role for TNF in the pathogenesis of cisplatin nephrotoxicity.

The biological activities of TNF are mediated by two functionally distinct receptors, TNFR1 (p55) and TNFR2 (p75). Many of the cytotoxic and proinflammatory actions of TNF are mediated by TNFR1 (22). However, studies in mice deficient in either TNFR1 or TNFR2 revealed that the nephrotoxic effects of cisplatin, at least those mediated by TNF, are signaled through TNFR2 rather than TNFR1 (23).

TNF is produced by many types of cells, but typically hematopoietic cells. To determine whether TNF was being produced by immune cells in cisplatin toxicity, we created chimeric mice in which the bone marrow was ablated and replaced with donor bone marrow cells from either wild-type (WT) or from TNF knockout mice. Chimeras with kidneys of WT animals developed significant renal failure after cisplatin treatment regardless of the immune cell source. Chimeras with kidneys of TNF knockout mice showed significantly less renal dysfunction, renal histologic injury, and urine and serum TNF levels; again regardless of the immune cell source. These results indicate that a substantial portion of circulating and urinary TNF is derived from non-immune cells, probably renal epithelial cells themselves, after cisplatin administration.

TOLL-LIKE RECEPTOR 4

Toll-like receptors (TLRs) are a family of receptors positioned as a first line of innate defense by recognizing pathogen-associated molecular patterns as well as endogenous signals of tissue injury (24,25). The most extensively characterized TLR, TLR4, is the receptor for the endotoxin of gram-negative bacteria (26,27) In addition to bacterial endotoxin, TLR4 can also be activated by endogenous molecules or danger signals released during tissue injury (28,29). Using a similar strategy as for TNF, we found that TLR4 is required for TNF production and subsequent nephrotoxicity after cisplatin injection and that this effect is also due to TLR4 expression on parenchymal cells rather than immune cells (30). These results suggest that renal cells may be both the sensors of injury and producers of cytokines which promote further injury

DENDRITIC CELLS

Dendritic cells are a key component of the innate immune response by acting as sensors of infection or injury and activating downstream events, including adaptive immune responses. Dendritic cells are abundant in the kidney, but their function in kidney health and disease is not well understood. Dendritic cells can have both pro-inflammatory and anti-inflammatory actions.

To evaluate the role of dendritic cells in cisplatin nephrotoxicity, we used a transgenic mouse model in which expression of the diphtheria toxin receptor (DTR) is driven by the CD11c dendritic cell promoter (31). Expression of the DTR renders these murine cells sensitive to the cytotoxic actions of diphtheria toxin (DT) such that the injection of DT leads to prompt and dramatic depletion of dendritic cells in both the kidney and spleen (32). Mice were depleted of dendritic cells by DT injection and then 24 hours later treated with cisplatin. Somewhat surprisingly, mice depleted of dendritic cells before or coincident with cisplatin treatment, but not at later stages, experienced more severe renal dysfunction, tubular injury, neutrophil infiltration, and greater mortality than nondepleted mice (32). Studies involving mixed bone marrow chimeras have shown that the worsening of cisplatin nephrotoxicity in dendritic cell–depleted mice was not a result of the dying or dead dendritic cells themselves. These results showed that resident dendritic cells reduce cisplatin nephrotoxicity and its associated inflammation.

One candidate for mediating the protective effects of dendritic cells is the anti-inflammatory cytokine, interleukin 10 (IL-10). We investigated the role of endogenous IL-10 in cisplatin nephrotoxicity using IL-10 knockout mice. Much like the dendritic cell depleted mice, IL-10 knockout mice sustained more severe kidney injury than WT mice, indicating that IL-10 serves an endogenous protective role (33). To determine if IL-10 production accounts for the protective effects of dendritic cells, chimeric mice were produced with a mixture of IL-10–deficient and DTR bone marrow cells. When the resulting mice were treated with DT, the IL-10–producing dendritic cells which express the DTR were ablated, leaving behind only the IL-10–deficient dendritic cells. The results of these studies revealed that IL-10 production by dendritic cells was responsible for a portion of the protective effects of dendritic cells (33).

NEUTROPHILS

Cisplatin administration causes an increase in kidney neutrophil content (17,30,32,34,35). Moreover, maneuvers which decrease cisplatin nephrotoxicity, such as inhibition of TNF or TLR4 signaling (17,23,30,36), inhibition of intercellular adhesion molecule 1 (34) or administration of IL-10 (37), are associated with a decrease in renal neutrophil content. However, studies from our lab (38) and others (35) have shown that depletion of neutrophils using an antineutrophil antibody had no effect on cisplatin-induced renal dysfunction or tubular necrosis even though renal neutrophil infiltration was effectively abolished. These results suggest that infiltrating neutrophils are not essential for cisplatin-induced renal injury and may be a reflection of the severity of injury rather than its cause. In this respect, cisplatin-induced AKI differs from ischemic AKI, in which a pathogenic role for neutrophils has been established. Among other mechanisms, in ischemic AKI neutrophils may form neutrophil extracellular traps in which nuclear DNA is extravasated from neutrophils and provokes injury. The enzyme peptidyl arginine deiminase 4 (PAD4) is necessary for neutrophils to produce neutrophil extracellular traps. When we depleted PAD4 from mice, the severity of ischemic AKI was markedly reduced (39), whereas depletion of PAD4 had no effect on cisplatin AKI (unpublished results).

SUMMARY

Cisplatin-induced AKI is an inflammatory condition in which AKI produces inflammation which propagates further kidney injury. Moreover, renal parenchymal cells are an important element of this inflammatory process, by sensing injury via TLR4 and then producing a variety of inflammatory chemokines and cytokines. The exact identity of the renal cells which do this is still under investigation. There are endogenous protective mechanisms which help limit kidney injury, including resident dendritic cells and IL-10. Although inflammation is important in these models of kidney injury, the clinical value of anti-inflammatory strategies in AKI remains unproven.

DISCUSSION

Mackowiak, Baltimore: Is cisplatin acute kidney injury extrapolatable to acute kidney injury in general? What are the pluses and minuses associated with that model?

Reeves, San Antonio: We've tested some of these pathways and other people have tested these pathways in other forms of AKI — like ischemic reperfusion injury or sepsis models — and some of them are shared, but some are distinct. I didn't have a chance to go into it, but we've looked for instance at the role of neutrophils in cisplatin toxicity, and it seems like they don't play a role — at least in our hands — whereas in ischemia reperfusion they do play a role. So, not all pathways are identical.

Lippman, Washington, DC: A really lovely talk. I was wondering if the elaboration of DAMPs might suggest that another receptor involved in this is RAGE, the receptor for advanced glycation end-products. The reason I particularly suggest this is because of the increased risk in diabetics when they get cisplatin. Also, because there are RAGE knockouts available that you can do this modeling and there are very readily available RAGE antagonists which could be clinically applicable in this model and maybe worth trying.

Reeves, San Antonio: Yes, that's a good point. We have not looked at RAGE. We actually have looked at the interaction of diabetes and some of these pathways. In tumor necrosis factor knockout mice it seems that that abrogates some of the increase in risk that diabetes produces, at least for ischemia-reperfusion models of AKI. Good point.

Humes, Boston: It seems to me that there are a lot of anti-inflammatory drugs in clinical trials for many other things. So, in this population particularly, it seems like it's a ready-made group to look at anti-TNF in platinum therapies. And, of course, that's in spite of Abul Abbas' talk where you would imagine there could be some consequences of modulating the proinflammatory immune system. But are there any trials?

Reeves, San Antonio: No. In fact, I just looked in clinicaltrials.org and there are no trials of TNF antagonists. I had some conversations with somebody a few years ago — an oncologist — about that, but they hadn't pursued it. I think in this setting of cisplatin — obviously the only people who are getting cisplatin are people that have cancer. So, if we mitigate the acute kidney injury, but make the cancer worse, that's not an acceptable strategy either. So, that's just the caveat. Although, some people have raised concerns that because the name tumor necrosis factor came from the observation that, if you inject it, it causes necrosis of tumors, that inhibiting it would be bad. There's actually some good evidence that inhibiting TNF may be beneficial for treating some tumors. But no, there hasn't been a trial that I am aware of.

Auerwater, Baltimore: Wonderful talk. Typically, all the ATN issues are associated with renally excreted drugs. At least that's how I've always thought of them. So, why doesn't it occur with adequately metabolized drugs that also I assume are exposed to kidney parenchyma?

Reeves, San Antonio: I don't know. It may be the levels. In the case of cisplatin there's accumulation in the kidney, and so the levels in the tissue are higher than the circulating levels.