Vancomycin-Induced Kidney Injury: Animal Models of Toxicodynamics, Mechanisms of Injury, Human Translation, and Potential Strategies for Prevention

Abstract

Vancomycin is a recommended therapy in multiple national guidelines. Despite the common use, there is a poor understanding of the mechanistic drivers and potential modifiers of vancomycin-mediated kidney injury. In this review, historic and contemporary rates of vancomycin-induced kidney injury (VIKI) are described, and toxicodynamic models and mechanisms of toxicity from preclinical studies are reviewed. Aside from known clinical covariates that worsen VIKI, preclinical models have demonstrated that various factors impact VIKI, including dose, route of administration, and thresholds for pharmacokinetic parameters. The degree of acute kidney injury (AKI) is greatest with the intravenous route and higher doses that produce larger maximal concentrations and areas under the concentration curve. Troughs (i.e., minimum concentrations) have less of an impact. Mechanistically, preclinical studies have identified that VIKI is a result of drug accumulation in proximal tubule cells, which triggers cellular oxidative stress and apoptosis. Yet, there are several gaps in the knowledge that may represent viable targets to make vancomycin therapy less toxic. Potential strategies include prolonging infusions and lowering maximal concentrations, administration of antioxidants, administering agents that decrease cellular accumulation, and reformulating vancomycin to alter the renal clearance mechanism. Based on preclinical models and mechanisms of toxicity, we propose potential strategies to lessen VIKI.

Vancomycin is a recommended therapy in multiple national guidelines,^1–6 with a dedicated guideline for dosing and monitoring strategies designed to maximize efficacy and minimize vancomycin-induced kidney injury (VIKI). Despite the common use, there is a poor understanding of the drivers and modifiers of VIKI. Several mechanisms underlying the nephrotoxicity of vancomycin have been identified; however, this has thus far not been translated into safer therapy for patients. Polypharmacy with other nephrotoxins is a major reason for compounded toxicity, and “avoiding nephrotoxic agents” is simply not an option for most patients. One variable well within the control of the clinician is the overall exposure of vancomycin that their patients receive. Vancomycin pharmacokinetic (PK) parameters, such as maximal concentrations (Cmax) and area under the concentration-time curve (AUC), allow for the exposures at the individual patient level to be compared to outcomes of efficacy and toxicity. Understanding the therapeutic window (i.e., minimum threshold exposures that define efficacy and ceiling concentrations that result in toxicity without meaningful gains in efficacy) is important for safe and effective care of patients. In this review, we chronicle rates of VIKI against the backdrop of dosing and monitoring guidelines that have been employed through the years. We explore preclinical models evaluating VIKI and describe how these models can provide insight into relationships that are difficult to discern clinically. We specifically review the pharmacokinetic/toxicodynamic (PK/TD) relationship for vancomycin derived from preclinical studies, mechanisms of toxicity derived from cellular and animal studies, and describe surrogates that have been used and are emerging to define VIKI. Finally, we propose future strategies for the early detection of VIKI and explore the potential for rescue/pro-phylactic therapy.

History of Vancomycin-Induced Kidney Injury (VIKI)

Vancomycin is a glycopeptide antibiotic originally isolated from the soil bacteria Amycolatopsis orientalis,⁷ previously known as Streptomyces orientalis, and was approved by the United States Food and Drug Administration (FDA) for patient use in 1958.⁸ Vancomycin use remained low after FDA approval because of a variety of toxicities related to purity that included pyrogen reactions, cochleotoxicity, vestibulotoxicity, and phlebitis.⁹ These adverse effects were initially ascribed to impurities in the formulation¹⁰ that was nicknamed “Mississippi mud,” because of its muddy brown look.¹¹ However, the rise of methicillin-resistant Staphylococcus aureus infections (MRSA)¹² led to increased vancomycin use. By the 1980s, vancomycin with a purity of 90% to 95% was available,¹⁰ and acute kidney injury (AKI) was relatively infrequent at approximately 5% of patients treated with vancomycin.¹³ Dosing guidelines from the 1980s until 2008 recommended vancomycin trough concentrations between 5 and 15 μg/ml.¹⁴ With AKI concerns very low and uncertainties about target concentrations being reached, multiple groups (including our own) recommended higher vancomycin exposures to maximize effi_cacy.^14–16

A standard for relating vancomycin use to AKI was created in the 2009 vancomycin therapeutic guidelines and was based on serial measurements of serum creatinine (an increase of 0.5 mg/dl or > 50% increase from baseline).¹⁴ Indeed after a definition was created in 2009 and the pneumonia¹⁷ and vancomycin guidelines¹⁴ recommended vancomycin trough concentrations of 15 to 20 μg/ml for serious infection, reports of kidney injury increased (Figure 1). In critically ill patients where higher vancomycin exposures were sought, rates of kidney injury approached 30%.¹⁸ From the clinical data, it also became clear that numerous factors amplified the risk for VIKI,¹⁹ including receipt of concomitant nephrotoxins, larger vancomycin doses (e.g., 4 g/day),²⁰ total duration of therapy,²¹ elevated trough concentrations (e.g., above ~ 15 μg/ml), infusion type,²² patient severity of ill-ness, pre-existing renal disease, and obesity. The interested reader is referred to well-written reviews that have summarized the clinical data.^{19, 23, 24} Since those reviews were composed, the latest and most comprehensive prospective study identified overall AKI rates of 26% among patients infected with MRSA bacteremia who were treated with vancomycin.²⁵ From this work, it became apparent that the exposure-efficacy curve was relatively flat with efficacy obtained with vancomycin AUCs as low as 94 mg/L × 24 hours. Maximal efficacy occurred at an upper threshold of 515 mg/L × 24 hours, whereas toxicity followed a linear trend across the ordinal categories (Figure 2). It should be noted that more work is needed to fully define the therapeutic window as prospective studies evaluating vancomycin efficacy and toxicity are limited. Contemporary strategies will focus on achieving enough vancomycin exposure to ensure efficacy, but limiting exposure to ensure that excess AKI does not occur. This review will focus on the mechanisms of VIKI and potential ways to further decrease the inherent risk of utilizing an antibiotic where exposure-toxicity and exposure-efficacy curves overlap.

Figure 1.

Number of citations in PubMed mentioning vancomycin and nephrotoxicity between 1977 and 2019. The pneumonia guidelines released in 2005 (highlighted in green) and vancomycin guidelines released in 2009 (highlighted in pink) both recommended higher trough concentrations.

Figure 2.

Quintile of Exposure by AUC: (1) 94–387, (2) 392–515, (3) 516–621, (4) 621–757, (5) 758–1755.²⁵ AUC, area under the curve; AKI, acute kidney injury.

Animal Models and Toxicodynamics

Despite more than 60 years of experience, the drivers of VIKI are not yet fully understood.²⁶ Understanding VIKI in the clinical arena is challenged by the high cost of conducting clinical trials, which have been estimated to exceed a median of $40,000 per enrolled patient.²⁷ For drugs like vancomycin that are already approved and in use, this financial barrier relegates most investigations to retrospective study. Retrospective analyses are subject to many sources of bias, and obtaining patients that are homogenous enough to detect differences in outcomes without confounding (or enrolling a large enough group to control and mitigate confounding) is challenging. Animal models, while surrogates of human physiology and subject to their own limitations, allow for the careful isolation of single variables and direct comparison of treatments. Additionally, animal models allow for the capture of copious amounts of data and samples that are only available clinically under the conduct of large-scale prospective trials. The FDA advises that acute toxicity studies are carried out to test the short-term adverse effects of a drug when administered either in single or multiple doses during a 24-hour period in two mammalian species (including one non-rodent) to identify species differences in toxicity.²⁸ Chronic toxicity studies are carried out to test adverse effects occurring after the repeated or continuous administration of a drug for up to 6 months.²⁹

Acute, chronic, and developmental toxicity studies of vancomycin have been conducted in experimental animals including mice, rats, dogs, cats, rhesus monkeys,³⁰ and rabbits. ³¹ In this review, we focus primarily on toxicity studies in rats and dogs as they are the most common and are more relevant for human translation as described below. To allometrically scale doses for the rat and dog, rat doses should be divided by 6.2 and dog doses should be divided by 1.8.³² This normalizes the total exposures between humans and the animal model of toxicity by matching AUCs between the various species (i.e., accounting for more rapid clearance of smaller mammals).

Rats

Routes of Administration and Nephrotoxicity

Acute toxicity studies conducted in the 1950s demonstrated that rats died in clonic convulsions immediately following intravenous administration of vancomycin (LD₅₀ 319 mg/kg; lethal dose 50% is the amount of the substance required to kill 50% of the test population), likely as a function of central nervous system toxicity. However at this point in history, the product was not yet purified, and impurities may have driven this adverse event^{10, 33} Intraperitoneal administration of vancomycin was better tolerated and high doses were needed to cause lethal events (LD₅₀ 2218 mg/kg).³³ Later in the 1980s when doses of 75, 150, and 350 mg/kg were given intraperitoneally twice daily for 4 days, the nephrotoxic dose of vancomycin was determined to lie between 150 to 350 mg/kg.³⁰

Studies have also been performed where vancomycin was administered via the subcutaneous route. It has been previously demonstrated that the subcutaneous route results in low levels of absorption, even in chronic toxicity studies with daily subcutaneous doses of 400 mg/kg or less for 7 months.³³ Large doses of 400 mg/kg/day given subcutaneously were not well tolerated and caused considerable necrosis of the skin at the site of injection³⁴ due to the acidic pH of the vancomycin solution.³⁵ Even with low absorption, a linear relationship was observed between subcutaneously administered vancomycin dose and vancomycin concentration in renal parenchymal tissue that ranged from 26 to 614 μg/g in rats receiving 10 to 400 mg/kg/day. Thus although studies employing the subcutaneous administration of vancomycin have occasionally induced histopathologic changes to the kidney tissue at high doses, low absorption from this route of administration and the lack of studies measuring blood concentrations preclude many conclusions.³⁴

It was further identified that exposure-toxicity relationships with vancomycin may be dependent upon dosing route for rat studies.³⁶ These authors demonstrated that vancomycin nephrotoxicity was highest with intravenous administration, followed by intraperitoneal, and lowest after intramuscular administration. A 500 mg/kg iv dose was lethal following the first, second, or third injection. This study suggested that absorption from the site of administration directly impacted exposure-toxicity relationships. Further, it showed that vancomycin has a high affinity for renal tissue. Following a single intravenous dose, vancomycin serum concentrations decreased with a half-life of 0.6 hours, but vancomycin kidney concentrations remained high for several days and accumulated in renal tissue after repeated dosing. A reduction of the nephrotoxic effect was observed when vancomycin was combined with D-glucaro-1.5-lactam, a β-glucuronidase inhibitor, which suggested that the nephrotoxic effect of vancomycin may be activated by lysosomal enzymes in the kidney cells.

Dose Effect and Nephrotoxicity

Dose-ranging studies revealed that an increase in the vancomycin dose (up to 450 mg/kg intraperitoneally) and duration of treatment (up to 28 days, subacute) in rats is associated with increased histopathologic damage and elevations in urinary biomarkers of AKI.^{37, 38} Damage is most prevalent at the proximal tubule, which is further supported by urinary biomarkers, such as kidney injury molecule-1 (KIM-1), clusterin, and osteopontin (OPN). Biomarkers KIM-1 and clusterin were the best predictors of histopathologic damage within 24 hours in the intraperitoneally dosed rat model of VIKI.³⁹

Pharmacokinetic/Toxicodynamic (PK/TD) Relationships: VIKI

The rat model has also been used to explore PK/TD relationships. Since vancomycin efficacy correlates with exposure relative to the minimum inhibitory concentration (AUC:MIC),⁴⁰ and since AUC remains constant with total daily dose, the authors asked whether the nephrotoxic potential of vancomycin (400 mg/kg/day intraperitoneally for 7 days) could be reduced if vancomycin was administered in divided doses (200 mg/kg/day twice daily intraperitoneally for 7 days).⁴¹ Although vancomycin concentrations were not measured, dose fractionation resulted in decreased kidney injury. Intravenous studies, where careful control of exposures is possible and exposures are not compromised by variable absorption rates, have been used to probe the PK parameters that cause toxicity and quantitative thresholds for injury. Rat studies employing intravenous dosing have demonstrated that vancomycin AUC_0–24 or Cmax_0–24, rather than minimum concentration (Cmin), predict urinary biomarker response.^42–44 A threshold AUC_0–24 of 482.2 mg h/L has been identified and correlates directly with similar human exposure-response data. Thus, moving to AUC monitoring to prevent toxicity (such as suggested in the published guidelines⁴⁵) is supported by preclinical models. Future work is still needed to determine if minimizing Cmax while maintaining AUC is a beneficial strategy. Corroborating the work from this group,⁴¹ recent intravenous studies in rats have demonstrated lower urinary biomarkers in rats that received their daily vancomycin dose fractioned as opposed to consolidated (i.e., 3 to 4 times daily compared to once or twice daily). In this study, 24-hour AUCs were held functionally constant while Cmax was lowered. The Cmax better described the kidney injury as defined by KIM-1 urinary concentrations when compared to AUCs (Akaike information criterion of −5.28 vs −1.95)⁴⁶; however, additional work is needed to further delineate this relationship and determine if altering administration schemes (e.g., continuous infusions) can lessen VIKI by decreasing Cmax even for similar AUCs.

Combination Therapy and Nephrotoxicity

Finally, it is well known in the preclinical literature that additional nephrotoxic agents increase vancomycin-mediated nephrotoxicity in an additive or synergistic fashion. Rat studies evaluating toxicity of combination therapy have demonstrated that concurrent use of vancomycin with aminoglycosides, including gentamicin, tobramycin, and arbekacin, potentiates nephrotoxicity^{30, 35, 47, 48} and should be avoided when clinically possible. Human studies have likewise confirmed synergistic toxicity between vancomycin and aminoglyco-sides.^{49, 50} We are unaware of preclinical data evaluating the combination of vancomycin with plazomicin, sisomicin, dibekacin, TS2037, and two novel arbekacin derivatives.

Fetal VIKI

The impact of vancomycin on the kidneys of the fetus has been studied. The effect of vancomycin on the developing rat fetus has been examined in developmental toxicology studies.³¹ No adverse effects were noted on fetal viability, weight, or morphology that included external anatomical and internal visceral anomalies and skeletal examination, at vancomycin doses up to 200 mg/kg; though, this dose was nephrotoxic to dams. A more recent study⁵¹ assessing KIM-1, a sensitive biomarker for VIKI,³⁹ demonstrated that pups from dams dosed during the first trimester had higher concentration of KIM-1 in the fetal kidneys compared to the second and third trimesters. This suggests an inverse relationship between trimester of exposure and kidney damage, which warrants further examination.⁵¹

Dogs

Kidney toxicity in dogs is comparable to that observed in rats. In early dose-escalation acute toxicity studies, dogs died several days following vancomycin dosing (intravenous LD₅₀ 292 mg/kg) due to renal failure (blood urea nitrogen (BUN) 250–300 mg/dl) rather than to effects on the central nervous system. Similar to rat data, when vancomycin was administered to dogs as a rapid intravenous injection, it caused a drop in blood pressure that was mediated through a histaminergic response.³⁰ In chronic toxicity studies with vancomycin in dogs, 19 of 22 dogs tolerated a 50 mg/kg dose given iv daily for 311 days; slight renal damage was observed in four of the 22 dogs. Emesis was frequent with loss of weight in some dogs.^{30, 33} More recently, a retrospective study of 29 dogs treated by a veterinary clinic concluded that vancomycin was reasonably tolerated with an all-cause AKI rate of 17% when given in clinical doses with therapeutic intent. The most common dose was 15 mg/kg intravenously every 6 hours, and all dogs received at least one additional antibiotic during the course of their therapy.⁵²

Mechanisms of VIKI

There is some consensus that vancomycin-induced renal damage is caused by intracellular accumulation of the drug,³⁶ though the exact method of vancomycin transport in proximal tubule cells has not been fully clarified. Vancomycin is excreted via glomerular filtration^{53, 54} and tubular secretion. Vancomycin is actively transported across the basolateral membrane into proximal tubule cells via the organic cation transporter system^{55, 56} (Figure 3). The visualization of large amounts of vancomycin at the brush border also points to either ongoing secretion or reabsorption of the drug at the apical side (i.e., from the urine).^{57, 58} In support, recent preclinical data have demonstrated vancomycin transport across the brush border membrane by apical endocytosis via dehydropeptidase⁵⁹ and megalin⁶⁰ and vancomycin-mediated inhibition of the expression and function of P-glycoprotein, an efflux transporter.⁶¹ Although mechanistically plausible, the involvement of the multidrug and toxin extrusion (MATE1 and MATE2-K) family of proteins in the secretion of vancomycin has yet to be demonstrated.

Figure 3.

Mechanisms of vancomycin-induced nephrotoxicity. (a) AIN: Vancomycin-induced AIN is caused by an immunologic response to vancomycin with interstitial infiltrate consisting of eosinophils, mast cells, plasma cells, lymphocytes, and macrophages. (b) ATN: (i) Oxidative stress: Vancomycin stimulates oxidative phosphorylation, as measured by an increase in oxygen consumption and an increase in ATP concentrations, which generate free radicals via formation of an iron complex. Free radicals induce lipid peroxidation leading to mitochondrial membrane depolarization and release of cytochrome C with activation of caspases leading to apoptosis. High levels of mitochondrial ROS can promote DNA damage, which in turn is obligatory for activation of PARP-1, an enzyme involved in DNA repair. Overactivation of PARP-1 activity leads to depletion of NAD⁺/ATP and cell death by necrosis. High doses of vancomycin upregulate the transcriptional expression of the Hmox1 gene, an indicator of cellular oxidative stress, and downregulate the expression of Cat, Gpx6, Gstk1, Sod2, and Sod3, which encode some of the main cellular antioxidants. Methyl-CPG-binding domain proteins are protein readers of methylation. In VIKI, MBD2 binds to the methylated CpG elements of miR-301a-5p promoter and activates the promoter, thus increasing transcription of miR-301a-5p. miR-301a-5p upregulates p53, which in turn induces caspase activity leading to apoptosis via miR-192–5p. Receptor-mediated endocytosis of vancomycin from the urine and transporter-mediated secretion of vancomycin from the blood into tubular epithelial cells causes entrapment of vancomycin in endosomes. Endosomes fuse with lysosomes, which leads to accumulation of vancomycin in lysosomes. This accumulation of vancomycin in lysosomes triggers a massive autophagy that is enhanced by suppressing mTOR activation and mediated by the activation of PKCδ through the direct interaction of Atg7 and PKCδ. In addition, the production of ROS may induce the increased permeability of lysosomes that are near mitochondria. Lysosomal membrane damage releases proteases including cathepsins into the cytosol that activate apoptotic effectors such as the mitochondria and caspases. Complete disruption of lysosomes may induce cytosolic acidification, which in turn provokes uncontrolled cell death by necrosis. (b) (ii) Cast formation: Nanospheric vancomycin aggregates interact with uromodulin and obstruct the tubular lumen. This leads to reduced renal blood flow via tubuloglomerular feedback mechanisms with a resultant decline in renal function. Various factors could lead to vancomycin supersaturation and precipitation within the tubular lumen including high concentrations and low urinary pH. Potential strategies for prevention. A: antioxidants (e.g., vitamins C and E); B: transport inhibitors (e.g., cilastatin, JBP485, fosfomycin); C: lysosomal membrane stabilizers (e.g., fosfomycin), lysosomal enzyme inhibitors (e.g., D-glucaro-1.5-lactam); D: autophagy inhibitors (e.g. chloroquine, bafilomycin A1), mTOR activators (e.g., MHY1485), PKCδ inhibitors (e.g., rottlerin); E: PARP-1 inhibitors (e.g., 3-aminobenzamide, olaparib, rucaparib, niraparib, talazoparib, veliparib, pamiparib, CEP 9722, E7016); F: corticosteroids. Abbreviations: AIN, acute interstitial nephritis; ATN, acute tubular necrosis; Atg, autophagy-related gene; ATP, adenosine triphosphate; Cat, catalase; DHP-1, dehydropeptidase-1; Gpx6, glutathione peroxidase 6; Gstk1, glutathione S-transferase kappa 1; Hmox1, heme oxygenase 1; LMP, lysosomal membrane permeabilization; MBD2, methyl-CpG-binding domain protein 2; miR, microRNA; NAD, nicotinamide adenine dinucleotide; mTOR, mammalian target of rapamycin; O2, oxygen; OAT1/3, organic anion transporter 1/3; OCT-2, organic cation transporter 2; PARP-1, poly (adenosine diphosphate ribose) polymerase 1; PEG, polyethylene glycol; Pgp, P-glycoprotein; PKCδ, protein kinase C delta; ROS, reactive oxygen species, SOD, superoxide dismutase; VIKI, vancomycin-induced kidney injury.

The mechanisms that underlie the pathogenesis of VIKI are multifactorial (Figure 3).

Acute Interstitial Nephritis

Vancomycin has been observed with rare cases of acute interstitial nephritis (AIN).^62–68 Vancomycin-induced AIN has been suggested to be cellular-mediated hypersensitivity⁶⁶ (Figure 3a). Interstitial edema has been reported on renal biopsy with infiltrate consisting of eosinophils, mast cells, plasma cells, lymphocytes, and macrophages.^{67, 68} The mechanism of vancomycin-induced AIN is not known.

It has been suggested that VIKI may be mediated by complement activation followed by development of an inflammatory response. In their toxicogenomic study, kidneys of mice receiving high-dose vancomycin showed changes in the expression of several transcripts from the complement pathway, including C3 and C4b, and Cxcl1, a biomarker of ischemic acute kidney failure that is produced in a complement-dependent manner.⁶⁹ The role of the complement pathway in VIKI needs to be elucidated.

Acute Tubular Necrosis

Vancomycin causes dose-dependent ATN,^70–75 which is attributed to oxidative stress in proximal tubule cells, autophagy, and obstructive cast formation^{76, 77} (Figure 3).

Oxidative Stress

Oxidative stress is an imbalance between free radicals and antioxidants within cells that leads to mitochondrial dysfunction and cellular apoptosis. Vancomycin stimulates oxygen consumption and causes a dose-dependent increase in ATP concentrations in cultured renal cells,⁷⁷ which indicates that vancomycin can stimulate mitochondrial oxidative phosphorylation.⁷⁸ Oxygen consumption can lead to free radical generation,⁷⁹ and free radicals are involved in the pathogenesis of VIKI through the inhibition of superoxide dismutase.⁷⁸ Since 2,3-dihydroxybenzoic acid, an iron chelator, could protect against VIKI, this was evidence that VIKI is mediated at least partly through oxidative stress via formation of an iron complex that catalyzes free radical formation, specifically the hydroxyl radical, since iron is crucial in the generation of this radical.⁸⁰ Free radicals can induce lipid peroxidation, a chain reaction influencing unsaturated fatty acids in cell membranes, leading to their damage, and production of malondialdehyde.⁸¹ As such, an increase in malondialdehyde levels and a decrease in antioxidative enzyme activities – reduced glutathione peroxidase and superoxide dismutase – as indicators of oxidative stress have been reported in the kidney tissue of rats treated with vancomycin; antioxidants, including a superoxide dismutase conjugate,⁷⁸ reversed or prevented these effects and downstream histopathologic damage. In addition, the superoxides produced cause mitochondrial membrane depolarization and subsequent release of cytochrome C with activation of downstream caspases^{59, 82}, ultimately leading to cell apoptosis.

Of note, superoxide anions are endogenous inducers of DNA single-strand breaks which is obligatory for activation of poly (adenosine diphosphate ribose) polymerase 1 (PARP-1),⁸³ an enzyme involved in DNA repair. The enzyme PARP-1 uses NAD⁺ as a substrate in the repair process, subsequently, cells utilize ATP to regenerate NAD⁺ stores. With high amounts of DNA damage, PARP-1 overactivation leads to depletion of NAD⁺/ATP and cell death by necrosis. Further support for the oxidative stress mechanism in VIKI is demonstrated by the overactivation of PARP-1 activity in rats administered vancomycin and attenuation of the renal injury on treatment with 1,5-isoquinolinediol, a PARP inhibitor.⁸⁴

The occurrence of oxidative stress and mitochondrial damage in VIKI has been confirmed at the genomic level. In mice treated with high-dose vancomycin (i.e., seven daily intravenous doses of 200 and 400 mg/kg vancomycin and seven daily intraperitoneal doses of 400 mg/kg vancomycin), the transcriptional expression of Hmox1, an indicator of cellular oxidative stress, was upregulated and the expression of Cat, Gpx6, Gstk1, Sod2, and Sod3, which encode some of the main cellular antioxidants, were downregulated.⁶⁹ The miRNA appears to be an important part of this VIKI mechanistic pathway. The miRNAs miR-301a-5p and miR-192–5p have been implicated in the development of VIKI through methyl-CpG-binding domain protein 2 (MBD2) and p53.^{85, 86} Methyl-CpG-binding domain proteins, are protein readers of methylation and are actively involved in DNA-methylation-mediated transcriptional regulation by binding to methylated CpG on the DNA and either blocking other protein factors from binding to DNA or by binding to promoters of actively transcribed genes.⁸⁷ In VIKI, MBD2 binds to the methylated CpG elements of miR-301a-5p promoter and activates the promoter, thus increasing transcription of miR-301a-5p, which upregulates p53, caspase activity, and vancomycin-induced apoptosis in HK-2 cells. Additionally, p53 has been shown to induce apoptosis via miR-192–5p in wild-type mice, which was suppressed in p53-KO mice.⁸⁶ Blocking miR-301a-5p⁸⁵ or inhibiting miR-192–5p⁸⁶ with respective antisense oligonucleotides ameliorated VIKI in C57BL/6 mice.

Receptor-mediated endocytosis of vancomycin from the urine and transporter-mediated secretion of vancomycin from the blood into tubular epithelial cells causes entrapment of vancomycin in endosomes. Endosomes fuse with lysosomes, which leads to accumulation of vancomycin in lysosomes. In 1992, vancomycin accumulation in the lysosomes of proximal tubule cells was first visualized using immunoelectron microscopy⁵⁸; and later confirmed by another group.⁵⁷ This accumulation of vancomycin in lysosomes may trigger a detrimental autophagy process. Autophagy is a lysosomal-dependent process of self-degradation of cellular components for the maintenance of cell homeostasis or in response to stress⁸⁸ and may serve as an adaptive and protective mechanism for cell survival. As such, autophagy in proximal tubules has been shown to protect against AKI.⁸⁹ However, massive autophagy and dysregulated autophagy may cause cell death.^{90, 91} Several autophagy-related genes (Atgs) or proteins have been identified that are required for the maturation of autophagosomes, deficiency of which can inhibit the autophagy process. Using a proximal tubule specific-Atg7-knockout mouse model, these authors succeeded in attenuating VIKI thus demonstrating a destructive role for autophagy in VIKI.⁹² Further, morphologic analysis of HK-2 cells showed that vancomycin-induced cell apoptosis was enhanced by rapamycin, an mTOR inhibitor, reduced by chloroquine, an autophagy inhibitor, and confirmed with a caspase activity assay. Using a coimmunoprecipitation and various Atg7 deletion constructs in renal cells treated with vancomycin, they concluded that Atg7 mediates renal tubule cell apoptosis through direct interaction and activation of PKCδ. Further research is necessary to confirm this molecular mechanism.

In addition, various stressors and the production of reactive oxygen species (ROS), may induce the permeabilization of lysosomes that are near mitochondria. Lysosomal membrane damage releases proteases including cathepsins into the cytosol that activate apoptotic effectors, such as the mitochondria and caspases. Complete disruption of lysosomes may induce cytosolic acidification, which in turn provokes uncontrolled cell death by necrosis.⁹³

Cast Formation

Cast formation may also exacerbate vancomycin-induced ATN. Uromodulin, the most abundant urinary protein, is exclusively produced by renal epithelial cells where it regulates salt transport, protects against urinary tract infection and kidney stones, and has immunomodulatory functions. In the tubular lumen, uromodulin forms filaments that constitute the matrix of hyaline casts.⁹⁴ Uromodulin has been shown to interact with vancomycin aggregates resulting in intratubular cast formation, inflammation, and ATN.⁷⁶ These authors detected vancomycin within uromodulin casts in nine patients with ATN and replicated the pathophysiology in a mouse model. Various factors could lead to vancomycin supersaturation and precipitation within the tubular lumen including high concentrations and low urinary pH.⁹⁵ In addition, vancomycin accumulation at the brush border has been observed. Using a monoclonal antibody to vancomycin with immunocytochemistry the presence of significant amounts of vancomycin was detected in the tubular lumen and in the microvilli, nuclei, and cytoplasm specifically in the S1 and S2 segments of the proximal tubules.⁵⁷ Nevertheless, further studies are warranted to determine whether vancomycin accumulation in the tubular lumen is a result of supersaturation and precipitation at urinary pH.

Human Translation

The current practice for classifying VIKI is based on the 2009 vancomycin guidelines.¹⁴ These guidelines suggest that nephrotoxicity is considered when at least two instances of elevated serum creatinine concentrations are observed in the setting of several days of vancomycin therapy. Elevated serum creatinine is defined as an increase of 0.5 mg/dl or greater than 50% increase from baseline, whichever is greater. Other common methods for classifying AKI are the Risk of renal dysfunction, Injury to the kidney, Failure of kidney function, Loss of kidney function, End-stage kidney disease (RIFLE),⁹⁶ Acute Kidney Injury Network (AKIN),⁹⁷ and Kidney Disease Improving Global Outcomes (KDIGO)⁹⁸ criteria (Table 1). These methods have similar discriminatory power for predicting mortality in critically ill patients.⁹⁹ The newest vancomycin guidelines (in draft form⁴⁵) report that most studies have assessed the relationship between vancomycin and nephrotoxicity using the aforementioned definition of a serum creatinine increase of 0.5 mg/dl or greater than 50% increase from baseline; however, some studies have begun to adapt a lower threshold (such as a serum creatinine absolute increase of ≥ 0.3 mg/dl), which qualifies patients as Stage 1 AKI by the AKIN and KDIGO criteria. The move to utilize absolute changes in creatinine as opposed to relative percentage changes is because the time to detection of renal insult is more heterogeneous across various stages of baseline kidney disease when using percentage changes as opposed to absolute changes.^{100, 101} That is, absolute changes, specifically when set to a lower value such as 0.3 mg/dl can identify an injury sooner and more consistently regardless of patient renal disease at baseline. The time to reach any set percentage rise in serum creatinine depends on the level of kidney dysfunction that a patient has prior to the AKI event. For instance in patients with an ultimate 50% reduction in creatinine clearance, patients with no chronic kidney disease (CKD) achieve a 100% increase in serum creatinine (SCr) within 12 hours, whereas those with stage 2 CKD approach the same doubling of SCr around 30 hours and those with stage 3 CKD do not double their SCr until after 60 hours.¹⁰²

Table 1.

Definitions and Classification/Staging Systems for Acute Kidney Injury^96–98

	Serum Creatinine	Urine Output	GFR
RIFLEa criteria
Risk	1.5 times baseline	< 0.5 ml/kg/hr for 6 hrs	↓ by > 25%
Injury	2 times baseline	< 0.5 ml/kg/hr for 12 hrs	↓ by > 50%
Failure	3 times baseline or ≥ 4 mg/dl with acute ↑ of ≥ 0.5 mg/dl	< 0.3 ml/kg/hr for 24 hrs or anuria for 12 hrs	↓ by ≥ 75%
Loss	Persistent failure = complete loss of function for > 4 weeks (requiring dialysis)
ESKD	Complete loss of kidney function for > 3 months (requiring dialysis)
AKIN criteria
Stage 1	↑ of ≥ 0.3 mg/dl within 48 hrs or 1.5–2.0 times baseline within 48 hrs	↓ to < 0.5 ml/kg/hr for > 6 hrs
Stage 2	2–3 times baseline	↓ to < 0.5 ml/kg/hr for > 12 hrs
Stage 3	> 3 times baseline or ≥ 4 mg/dl with acute ↑of ≥ 0.5 mg/dl	↓ to < 0.3 ml/kg/hr for 24 hrs or anuria for 12 hrs
KDIGO criteria
Stage 1	↑ of ≥ 0.3 mg/dl within 48 hrs or 1.5–1.9 times baseline within 7 days	< 0.5 ml/kg/hr for 6–12 hrs
Stage 2	2.0–2.9 times baseline	< 0.5 ml/kg/hr for ≥ 12 hrs	< 18 yrs, ↓ to
Stage 3	3 times baseline or ≥ 4 mg/dl or initiation of RRT	< 0.3 ml/kg/hr for ≥ 24 hrs or anuria for ≥ 12 hrs	< 35 ml/min/1.73 m2

GFR = glomerular filtration rate; RIFLE = Risk of renal dysfunction, Injury to the kidney, Failure of kidney function, Loss of kidney function, End-stage kidney disease; ESKD = end-stage kidney disease; AKIN = Acute Kidney Injury Network; KDIGO = Kidney Disease Improving Global Outcomes; RRT = renal replacement therapy.

^aClassification is based on the highest score for any category.

More work will be needed to define the level of clinical kidney injury that should be utilized for adverse event attribution (i.e., VIKI). Within the criteria to define AKI, serum creatinine is the most commonly applied methodology because it is commonly measured in practice and thus available for assessment in retrospective studies. Creatinine is known to be an insensitive and slowly reactive marker of kidney function (vis-à-vis extrapolated glomerular filtration rate).¹⁰³ Glomerular filtration rates can drop as much as 50% before changes in creatinine are detected.^104–106 Patient hydration status also can have a big effect on serum creatinine values.¹⁰⁷ Acute kidney injury-induced fluid retention and exogenous fluid administration can actually make serum creatinine values drop, delaying recognition of AKI. Therefore, there has been an invigorated interest in biomarkers that might detect injury closer to the actual event. Within 3 to 6 hours after an ischemic-reperfusion event, KIM-1 is elevated in experimental studies; whereas, creatinine elevations in this time period were less detectable and did not consistently elevate within the 120-hour study period.³⁷ Hence, understanding when the injurious event occurs might be more beneficial to the clinical care of the patient than utilizing a surrogate marker (e.g., SCr) of a functional clearance metric (e.g., glomerular filtration rate (GFR)). Future studies will be required to understand how biomarkers predict VIKI in humans; these studies will be facilitated once biomarkers are approved for clinical use. Presently, urinary biomarkers (e.g., KIM-1 and clusterin) are qualified for rat¹⁰⁸ by the FDA (i.e., for drug-induced AKI); a composite panel of six biomarkers that include KIM-1 and clusterin has been qualified in human drug trials.¹⁰⁹ These biomarkers (i.e., KIM-1 and clusterin) will likely have direct relevance to identifying and quantifying VIKI. An understanding of time course and magnitude of change will be needed before updated vancomycin guidelines can be developed to prevent additional injurious exposures.

Potential Strategies for Prevention

Understanding the mechanisms underlying the pathogenesis of VIKI has aided the design of strategies to prevent it (Table 2; Figure 3; Table S1); yet few of the preclinical studies that have shown promise have been implemented in clinical trials to date for VIKI. We are only aware of a single trial, namely NCT03921099, where 40 patients are to be enrolled to evaluate the impact of ascorbic acid in the prevention of vancomycin-induced nephrotoxicity.¹¹⁰

Table 2.

Summary of Studies That Have Evaluated Potential Strategies for Prevention of VIKI

	Compounds to Inhibit or Reverse VIKI	Models
Antioxidants
1	Hexamethylenediamine-conjugated superoxide dismutase (AH-SOD) 78	Wistar rats
2	Erdosteine122	Wistar albino rats
3	a-lipoic acid, ginkgo biloba extract, melatonin, or amrinone131	Sprague Dawley rats
4	Erythropoietin121	Wistar albino rats
5	Caffeic acid phenyl ester (CAPE), vitamin C, vitamin E, N-acetylcysteine 20	Wistar albino rats
6	2,3-dihydrobenzoic acid (DHB)80	Wistar albino rats
7	Curcumin139	Sprague Dawley rats
8	1,5-isoquinolinediol (ISO)84	Wistar rats
9	Thymoquinone140	Wistar albino rats
10	Atorvastatin119	Wistar albino rats
11	Spirulina, pycnogenol, spirulina + pycnogenol112	Wistar albino rats
12	Vitamin E, mitoTEMPO (mitochondria-targeted antioxidant), vitamin C, N-acetylcysteine, glutathione141	Porcine proximal tubular epithelial LLC-PK1 cell line
13	Naringenin113	Wistar albino rats
14	Zingerone114	Sprague Dawley rats
15	Docosahexaenoic acid-enriched phosphatidylcholine (DHA-PC)116	Balb/c mice
16	Rutin117	Wistar rats
17	Silymarin118	Wistar albino rats
18	Zingerone115	Sprague Dawley rats
Transport inhibitors
1	Imipenem-cilastatin, flomoxef sodium, fosfomycin sodium142	Rabbits
2	Cilastatin123	Rabbits
3	Cilastatin60	C57BL/6J mice
4	Cilastatin61	C57BL/6J mice
5	JBP485125	Wistar rats
Formulations
1	Polyethylene glycol (PEG)-vancomycin conjugate143	Bacterial efficacy studies
2	Vancomycin with 2 inactive ingredients, D-mannitol and Macrogol400 (PEG400), referred to as MEEK132,149	Sprague Dawley rats
3	Nanoconjugated vancomycin138	Swiss mice
4	Vancomycin PEGylated liposomes134	42 CF-1 mice
5	Dicetylphosphate (DCP) liposomal vancomycin144	CD-1 mice
6	Liposomal vancomycin145	Kunming mice
7	Vancomycin loaded pH-responsive chitosan nanoparticles containing anionic gemini surfactant (DL_CSSNPs)137	Mice
8	Vancomycin analog compounds146	Human renal proximal tubule epithelial cells (HK-2), human liver cell line (HL-7702)
9	Vancomycin loaded chitosan sponges136	Bacterial inhibition assay
10	Vancomycin-dextran sulfate sodium nanoplexes147	Adenocarcinoma human alveolar basal epithelial cells (A549), embryonic kidney cells (HEK-293), and liver hepatocellular carcinoma (Hep G2) cell lines
11	pH-responsive liposomal formulations DOAPA-VAN-Lipo, and DLAPA-VAN-Lipo135	Balb/c mice
12	Vancomycin loaded silver nanoparticles148	Bacterial efficacy studies
Others
1	Potassium salt of D-glucaro-1.5-lactam (a β-glucuronidase inhibitor)36	Wistar albino rats
2	Fosfomycin126	Wistar albino rats
3	Fosfomycin, imipenem/cilastatin127, 150	Wistar albino rats
4	Fosfomycin128	Wistar rats
5	Dexmedetomidine130	Wistar albino rats
6	Chloroquine92	Proximal tubule Atg7 wild type and deficient mice

AH-SOD = hexamethylenediamine-conjugated superoxide dismutase; Atg = Autophagy related gene; AUC = area under the concentration time curve; BUN = blood urea nitrogen; CAPE = caffeic acid phenyl ester; CMC = carboxy methyl cellulose; DCP = Dicetylphosphate; DL_CSSNPs = vancomycin loaded pH-responsive chitosan nanoparticles containing anionic gemini surfactant; DMSO = dimethylsulfoxide; DHA-PC = Docosahexaenoic acid-enriched phosphatidylcholine; DHB = 2,3-dihydrobenzoic acid; ERK = extracellular signal related kinase; i.m. = intramuscular; i.p. = intraperitoneal; ISO = isoquinolinediol; mTOR = mammalian target of rapamycin; NAR = naringenin; PEG = polyethylene glycol; PKCδ = protein kinase C delta; s.c. = subcutaneously; VIKI = vancomycin-induced kidney injury; VRSA = van-comycin resistant staphylococcus aureus; VSSA = vancomycin susceptible staphylococcus aureus.

Hydration

The impact of hydration on VIKI has been tested. Water 40 ml/kg given orally 30 minutes before and 30 minutes following a 350 mg/kg twice daily intraperitoneal dose of vancomycin appeared to reduce the degree of renal injury as evaluated by the reduction in BUN and creatinine values for rats treated with vancomycin compared with controls who did not receive vancomycin, however, it was not clear if this effect was secondary to hemodilution. Supporting that the effect may have been driven by hemodilution, biochemical indicators of nephrotoxicity including transport and metabolism functions in renal cortical slices, and relative kidney weights, were lower than controls not given vancomycin.³⁰ Thus blood dilution of the biomarkers may have been solely responsible for the differences. Additional work is needed to see if improving hydration status can help minimize vancomycin toxicity.

Antioxidants

As oxidative stress has been described as one of the mechanisms of VIKI (Figure 3), several antioxidants have been studied preclinically in an attempt to prevent VIKI. The effects of different antioxidants, such as α-lipoic acid, ginkgo biloba, melatonin, erdosteine, vitamin E, vitamin C, N-acetylcysteine, caffeic acid phenethyl ester, erythropoietin, 2,3-dihydroxybenzoic acid, tempol, thymoquinone, curcumin, atorvastatin, 1,5-isoquinolinediol and a superoxide dismutase conjugate, reversed or prevented oxidative stress and downstream histopathologic damage. These studies have been reviewed in detail.¹¹¹ Since then, spirulina and pycnogenol,¹¹² mitoTEMPO, naringenin,¹¹³ a free radical scavenger zingerone,^{114, 115} DHA-enriched phosphatidylcholine (DHA-PC),¹¹⁶ rutin,¹¹⁷ and silymarin¹¹⁸ have been evaluated in animal studies. In our opinion, the antioxidants that are most ready for clinical trials include preparations that are already clinically available, such as atorvastatin, vitamin C, vitamin E, N-acetylcysteine, erythropoietin, and erdosteine.

The data that underpin the potential role of the antioxidants in preventing VIKI follow. Atorvastatin provided a significant degree of protection in a rat model, but it depended on the dose of atorvastatin used: 5 mg/kg was insufficient, 10 mg/kg was most effective, and 20 mg/kg was toxic.¹¹⁹ The effects of vitamin C, vitamin E, and N-acetylcysteine on vancomycin-induced nephrotoxicity were investigated in a rat model.¹²⁰ The three antioxidants decreased the histologic damage caused by vancomycin in the kidney, but they did not completely prevent it. The vancomycin plus vitamins C and E groups showed mild epithelial desquamation, interstitial edema, and disarrangement. Vitamin E was the most effective in preventing vancomycin-induced tubular damage, followed by vitamin C and N-acetylcysteine.¹²⁰ Erythropoietin and erdosteine decreased histopathologic damage to rat kidneys.^{121, 122}

Transport Inhibitors

One strategy to minimizing VIKI is limiting the amount of cellular accumulation. A number of studies in mice and in primary porcine proximal tubule cells have demonstrated that cilastatin can decrease cellular vancomycin concentrations in the kidney and attenuate VIKI in a dose-dependent manner¹²³ by inhibiting the apical uptake of vancomycin through the megalin receptor,⁶⁰ the enzyme dehydropeptidase-I,⁵⁹ and by increasing P-glycoprotein expression and function.⁶¹ Although cilastatin might decrease the plasma concentration of vancomycin,¹²⁴ which is essential for efficacy.

It has been confirmed that the renal expression of P-glycoprotein is decreased with vancomycin along with other proximal tubule transporters – organic anion transporter (Oat) 1, Oat3, organic cation transporter 2 (Oct2), and another efflux protein, multidrug resistance-associated protein 2 (Mrp2). This decrease in transporter expression with vancomycin could lead to accumulation of endogenous endotoxins thus contributing to VIKI. Compound JBP485, a dipeptide isolated from the human placenta, demonstrated antioxidant and antiapoptotic effects and regulated the expression of the proximal tubule transporters to enhance renal excretion of endogenous uremic toxins and attenuate VIKI in rats without affecting plasma concentrations of vancomycin.¹²⁵

Miscellaneous Mechanisms

D-glucaro-1.5-lactam, β-glucuronidase inhibitor, has been reported to reduce the nephrotoxic effect of vancomycin in the rat as measured by impact of renal tubule cells,³⁶ presumably due to inhibition of lysosomal enzymes. The protective effects of fosfomycin against VIKI in rats have been studied by different groups.^126–128 Fosfomycin significantly decreased renal accumulation of vancomycin and reduced nephrotoxicity in a dose-dependent manner. The renal protective effects of fosfomycin on vancomycin-induced nephrotoxicity may be due to the inhibition of the tubular transport of vancomycin.¹²⁷ Of note, fosfomycin has been shown to protect the kidneys against dibekacin-induced nephrotoxicity in rats by stabilizing the lysosomal membrane.¹²⁹ Although this mechanism has not yet been demonstrated in VIKI, it is plausible. Another compound, an α−2 receptor agonist, dexmedetomidine, was shown to reduce the extent of vancomycin-induced renal damage in rats by preventing the elevation of vasoconstrictor agents.¹³⁰ Amrinone, a phosphodiesterase type III inhibitor,¹³¹ was effective in reducing VIKI in rats but may not find clinical application because of its cardiovascular side effects. Recently, the autophagy inhibitor chloroquine has been shown to protect against vancomycin-induced nephrotoxicity in rats.⁹² Whether the above-mentioned modalities are protective against VIKI in human subjects remains to be determined.

Formulation Alterations

Multiple attempts have been made to alter the vancomycin formulation to lessen VIKI, and various formulations are under development to improve the safety and efficacy of vancomycin. Liposomes and chitosans are among the best examples, though one miscellaneous formulation bears mention. The inactive ingredients D-mannitol and Macrogol 400 (polyethylene glycol 400 (PEG 400), in MEEK, a generic vancomycin formulation, reduced the nephrotoxicity of vancomycin in rats.¹³² It is unclear if this has clinical implication.

Liposomes

Liposomal encapsulation has been employed for multiple nephrotoxic drugs and has decreased kidney injury. Liposomal encapsulation enhances reticuloendothelial system uptake, consequently killing bacteria trapped within macrophages, and resulting in increased levels of vancomycin in spleen and liver. Liposomes deliver higher concentration of vancomycin to lung tissue while reducing accumulation in kidney tissue¹³³; surface PEGylation augments this effect.¹³⁴ Novel fatty acid-based zwitterionic lipids have been used to develop pH-responsive liposomes for the targeted delivery of vancomycin to acidic environments.¹³⁵ Bacteria thrive in these conditions whereas antibiotics often lose their activity. There is potential for liposomes to release vancomycin in response to change in pH at the site of infection. Reduction in kidney cellular concentrations may translate to safer therapies for patients, but these formulations need to be further developed.

Chitosans

A chitosan sponge as a carrier matrix for sustained vancomycin release was evaluated in vitro for its sustained-release effect.¹³⁶ The high hydrophilic nature of vancomycin was cited as the reason it did not produce the desired sustained release, but sustained-release formulations such as these could potentially reduce vancomycin nephrotoxicity if site-concentrated therapy is desired. In another study, vancomycin loaded pH-responsive chitosan nanoparticles for controlled and targeted delivery of vancomycin showed an increase in vancomycin release at pH 6.5 compared to pH 7.4 and reduced MRSA burden 8-fold compared to vancomycin.¹³⁷

Nanoconjugated vancomycin prepared from a carboxymethyl derivative of chitosan with folic acid showed beneficial effects against a lymphocyte model with S. aureus.¹³⁸

Summary

In summary, preclinical studies have demonstrated that VIKI is a result of drug accumulation in proximal tubule cells, specifically in the lysosomes. Vancomycin accumulation at the proximal tubule triggers degradative autophagy, oxidative stress, and apoptosis. Altered PK exposure and a variety of compounds may help to circumvent these mechanisms of toxicity. Future studies are needed to bridge preclinical knowledge to the bedside and determine if these strategies will result in safer vancomycin therapy for patients.

Supplementary Material

supplementary

Table S1. Summary of studies that have evaluated potential strategies for prevention of VIKI.