Adverse outcome pathway (AOP): α2u-globulin nephropathy and kidney tumors in male rats

Abstract

The National Research Council’s vision of using adverse outcome pathways (AOPs) as a framework to assist with toxicity assessment for regulatory requirements of chemical assessment has continued to gain traction since its release in 2007. The need to expand the AOP knowledge base has gained urgency, with the U.S. Environmental Protection Agency’s directive to eliminate reliance on animal toxicity testing by 2035. To meet these needs, our goal was to elucidate the AOP for male-rat-specific kidney cancer. Male-rat-specific kidney tumors occur through the ability of structurally diverse substances to induce α2u-globulin nephropathy (α2u-N), a well-studied mode of action (MoA) not relevant in humans that results in kidney tumor formation in male rats. An accepted AOP may help facilitate the differentiation from other kidney tumors MoAs. Following identification and review of relevant in vitro and in vivo literature, both the MIE and subsequent KEs were identified. Based on the weight of evidence from the various resources, the confidence in this AOP is high. Uses of this AOP include hazard identification, development of in vitro assays to determine if the MoA is through α2u-N and not relevant to humans resulting in decreased use of animals, and regulatory applications.

Graphical Abstract

Keywords:

• α2u-Globulin
• kidney
• nephropathy
• adverse outcome pathway
• molecular initiating event
• key event

Introduction

Regulatory drivers, such as the United States Environmental Protection Agency’s (US EPA) directive to eliminate reliance on animal toxicity testing by 2035 (US EPA 2019), and technological advances are modernizing toxicology testing and risk assessment. New approach methods (NAMs) that reduce reliance on animal testing either for chemical hazard identification or studies to define the mode of action (MoA) and human relevance are inherent to this process. Confidence in NAMs can be increased by documenting biological response pathways, called adverse outcome pathways (AOPs) that temporally and physiologically link molecular and cellular biological events to apical outcomes in animal studies. Internationally, efforts are underway to document a wide range of biological key events (KEs) in a standardized manner to move toward both a consistent scientific understanding of relationships between KEs and alignment on levels of confidence needed to apply AOPs in decision-making processes (Becker et al. 2015; Perkins et al. 2015; Edwards et al. 2016).

To contribute to this effort, we propose an AOP based on well-documented pathological observations of empirical evidence obtained from a diverse set of non-genotoxic chemicals that cause kidney tumors in male but not female, rats or mice and also induces pathological observations unique to male rats, referred to as α2u-globulin (α2u) nephropathy (α2u-N) (Swenberg et al. 1989; Borghoff et al. 1990; US EPA 1991; IARC 1999; Swenberg and Lehman-McKeeman 1999). These pathological observations, α2u-N, are characterized by the accumulation of α2u, a low-molecular-weight protein that is synthesized in the liver of male rats, in the kidney.

Although humans do not synthesize α2u, they do synthesize other low molecular weight proteins in the same lipocalin superfamily as α2u based on limited sequence homology, a similar molecular weight, and a proposed function of transport (Brooks 1987; Pervaiz and Brew 1987; Löbel et al. 2001). However, Borghoff and Lagarde (1993) demonstrated that humans do not possess a protein similar to a2u in their kidneys with respect to the relative abundance of this protein or ligand binding characteristics and therefore not be at risk of developing chemically induced protein-mediated nephropathy. Unlike male rats, female rats, NBR male rats, or mice do not synthesize α2u in the liver, which is the major source of this protein that is filtered at the glomerulus and contributes to the protein droplets observed in kidneys of untreated male rats. Rodents that do not synthesize a2u in their liver or synthesize a2u (i.e. mice) do not develop α2u-N or kidney tumors when exposed to the chemicals that cause these responses in the male rats (Short, Burnett, et al. 1989; Short, Steinhagen, et al. 1989; US EPA 1991; Dietrich and Swenberg 1992a, 1992b). Both the U.S. EPA (US EPA 1991) and the International Agency for Research on Cancer (IARC 1999) developed criteria for determining if a male rat kidney carcinogen operates through an α2u-N-specific MoA as described and outlined in these published documents.

Chemically-induced α2u-N that occurs in male rats is a well-delineated MoA that has been found to lead to kidney tumors in male rats with chronic exposure. This MoA is supported by multiple examples of non-genotoxic chemicals that have been shown to induce these kidney effects, defined by protein droplet formation in the P2 segment of the renal proximal tubules, epithelial cell shedding and cytolysis, regenerative cell proliferation, along with granular cast formation in the outer medulla, and linear mineralization in the papilla (Charbonneau et al. 1989; Borghoff et al. 1990, 2001, 2009). The objective of this evaluation is to characterize the evidence base for the α2u-N kidney tumor MoA in the framework of an AOP by identifying and providing support for, the molecular initiating event (MIE), key events (KEs), and key event relationships (KERs) based on the Organization for Economic Cooperation and Development (OECD) AOP User’s Handbook (OECD 2018). Building this AOP is a first step to expanding the AOP knowledge-based network to organize this well-recognized pathway in a framework that can subsequently be used to align on a consistent set of in vitro assays to determine the likelihood of a chemical to elicit the identified MIE and or subsequent KEs in the AOP.

Materials and methods

Targeted literature search and review

Consistent with the OECD AOP User’s Handbook (OECD 2018), the literature review incorporated elements of best practice for identification, evaluation, and prioritization of data sources. Additionally, data sources, including reviews (secondary literature) from authoritative bodies that describe in detail the MoA for male rat kidney tumors were used to determine the relevancy of the literature for consideration in the development of this AOP (US EPA 1991; IARC 1999).

As a starting point for the literature search, targeted literature searching was performed from PubMed on 1 October 2020 to identify key peer-reviewed publications using the terms “α2u,” “alpha2u,” α2u, or “α2u globulin.” The targeted literature search identified 116 articles that were screened by title and abstract to determine utility to confirm and establish the MIE, KEs, and KERs critical to the development of the AOP for kidney tumors that developed through the α2u-N MoA. The full-text literature identified to be potentially relevant, based on screening the title and abstract, was then critically reviewed and categorized using the available data to inform the MIE, KE, or KERs. While α2u-N has been studied for over 30 years, this male rat kidney tumor MoA has not been characterized in terms of an AOP framework in which the MIE and subsequent KEs associated with α2u-N and kidney tumors are identified.

Organization into the AOP framework

The AOP Framework and template, as described by the OECD’s AOP User’s Handbook (OECD 2018) were used to characterize, develop and review this AOP in a transparent and scientifically-based method. Each event (MIE, KE, and KER) is described in terms of both a general biological description and the specific measurement method(s). Based on the empirical evidence supporting each KE within the AOP, the domains of applicability are defined in the section entitled AOP assessment.

Assessment of relative confidence in the AOP

Recognizing that AOPs are living documents intended to be used in regulatory decision-making in a fit-for-purpose manner according to the relative confidence in the AOP (Edwards et al. 2016), this section summarizes the weight of the evidence (WOE) for this AOP through consideration of biological plausibility, empirical evidence supporting the KERs and essentiality of all KEs (Section AOP Assessment).

Defining questions for each component in the WOE were used to anchor the relative confidence for the overall conclusion for the AOP. Biological plausibility was assessed on the availability of a clear mechanistic relationship between upstream and downstream KEs; essentiality was assessed by confirming the prevention of a downstream KE after blocking an upstream KE, and empirical evidence was assessed on dose and temporal concordance across multiple KEs (OECD 2018). In alignment with the OECD guidance and due to the large number of chemicals associated with this MoA, and the fact that this MoA is recognized by regulatory agencies (US EPA 1991: IARC 1999), the quantitative data set was not reported in its entirety.

Results

AOP for kidney tumors in male rats

Based on data developed on structurally diverse chemicals, the MIE is identified as chemical bonding to the protein α2u, KE1 is the decreased rate of α2u catabolism leading to KE2, the exacerbation of α2u accumulation in the form of protein droplets within renal proximal tubular cells. Repeated exposure to chemicals that induce α2u-N leads to sustained cycles of cytotoxicity and compensatory proliferation in the proximal tubular cells (KE3). Continued, chronic exposure eventually leads to linear mineralization of the kidney papilla (associative event) and observations of atypical kidney tubular hyperplasia (KE4), in some cases progressing to a low incidence of kidney adenomas/carcinomas [adverse outcome (AO)] (Tables 1(A,B) and Figure 1). The MIE, KEs, and AO are described in detail below, followed by a description of the domain of applicability of this AOP and an overall assessment of the confidence of the AOP.

Figure 1. Empirical evidence supporting the AOP. Effect levels shown in gray indicate the highest no effect level per study; effect levels in orange indicate the lowest observed effect level per study. Route of exposure shown by shape: circle, inhalation; square, oral. Short-term exposure is defined as <28 days; mid-term exposure is >28 and <300 days; long-term exposure is >300 days. 1,3-dichlorobenzene has only a single data point and so is not included in the graph. Full study details can be found in Table 3. ETBE: ethyl tertiary butyl ether; IP: isophorone; MIBK: methyl isobutyl ketone; MTBE: methyl-tert butyl ether; TBA: tertiary butyl alcohol.

Table 1 (A). Events: molecular initiating event (MIE), key events (KEs), and adverse outcome (AO).

	Key events	Title	Strength of evidence
1	Molecular initiating event (MIE)	Binding to α2u (reversible)	High
2	Key event 1 (KE1)	Decreased rate of α2u catabolism (renal proximal tubular cell lysosomes)	Medium
3	Key event 2 (KE2)	Increased accumulation of α2u resulting in exacerbation of protein droplets (renal proximal tubular cells)	High
4	Key event 3 (KE3)	Cytotoxicity-induced cell proliferation (renal proximal tubular cells)	High
5	Key event 4 (KE4)	Atypical hyperplasia (renal tubules)	High
6	Adverse outcome (AO)	Renal adenomas/carcinomas	High

Table 1 (B). Relationships between two key events (including MIE and adverse outcome).

Key event relationship (KER)	Title	Adjacency^a	Strength of evidence
Key event relationship 1 (KER 1)	Binding to α2u leads to the decrease of α2u catabolism	Adjacent	High
Key event relationship 2 (KER 2)	Decreased rate of α2u catabolism leads to increased accumulation of α2u in the proximal tubule; exacerbation of protein droplets containing α2u	Adjacent	High
Key event relationship 3 (KER 3)	Exacerbation of α2u protein droplets leads to cytotoxicity and compensatory regenerative cell proliferation in the proximal tubule	Adjacent	High
Key event relationship 5 (KER 4)	Sustained cytotoxicity and compensatory regenerative cell proliferation leads to atypical hyperplasia in the renal tubules	Adjacent	High
Key event relationship 5(KER 5)	Increased, atypical hyperplasia, described by lesions specific to exacerbation in protein droplets (i.e. linear mineralization) (AE) to an increase in renal adenomas/carcinomas	Adjacent	High

^aAdjacency reflects the distance between the two KE described by the KER, where adjacent describes the relationship between two KEs that occur immediately upstream or downstream of each other in the AOP.

The molecular initiating event (MIE): binding to α2u protein

The MIE is the reversible binding of a chemical to the α2u protein. α2u is a low-molecular-weight protein (∼18 700 Da) that is synthesized in the liver and secreted at a rapid rate into the blood of male, but not female, rats, with ∼50% of the protein filtered at the glomerulus being resorbed into the proximal tubule cells (Roy and Raber 1972; Roy et al. 1976). Chemical binding to α2u can decrease its rate of catabolism, leading to its accumulation in the form of protein droplets in proximal tubule cells. A diverse group of chemicals, or their metabolites, have been reported to bind to α2u, either in vivo or in vitro including the following: 1,4-dichlorobenzene (-2,5-dichlorophenol, metabolite), d-limonene (d-limonene 2-5-oxide, metabolite), methyl isobutyl ketone (MIBK), 2,2,4-trimethylpentane, (2,4,4-trimethyl-2-pentanol, metabolite), tertiary butyl alcohol (TBA), ethyl tertiary butyl ether (ETBE) (TBA, metabolite), and methyl tert-butyl ether (MTBE) (TBA, metabolite) (Borghoff and Lagarde 1993; Swenberg 1993; Prescott-Matthews et al. 1999; Borghoff et al. 2015, 2017). Following exposure to one of these chemicals, either the parent compound or metabolite were found to co-elute with α2u when purified from the male rat kidney (Lock et al. 1987; Charbonneau et al. 1989; Prescott-Matthews et al. 1999). Borghoff et al. (1991) demonstrated chemical binding to α2u in vitro with several chemicals showing varying binding affinities that range from 10⁻⁷ to 10⁻⁴ M. Molecular modeling was then used to evaluate the critical components of these structures to bind to this protein, with the analysis demonstrating that binding is dependent on both hydrophobic interactions and hydrogen bonding. Examples of chemical stressors are provided in Table 2.

Table 2. Stressors (examples).

Chemical	Evidence^a,b
Methyl isobutyl ketone (MIBK)^a	High
Ethyl tertiary butyl ether (ETBE)—TBA (metabolite)^b	Medium
Methyl tert-butyl ether (MTBE)—TBA (metabolite)^a	High
Tertiary butyl alcohol (TBA)^a	High
2,4,4-Trimethylpentane-2,4,4-trimethyl-2-pentanol (TMPOH) (metabolite)^a	High
d-Limonene—d-limonene-oxide (metabolite)^a	H
1,4-Dichlorobenzene-2,5-dichlorophenol (metabolite)^a	High
Decalin^b	Medium
Isophorone^a	Medium

^aEvidence considered “High” if direct evidence is available for the MIE, KE2, KE3, and KE4. Direct evidence for KE1 is limited, but for many of the stressors, there is indirect evidence for KE1.

^bEvidence considered “Medium” if evidence was available for smaller subsets of KE (MIE, KE2 and KE4; KE2, KE3, and KE4; or KE3 and KE4).

Key event 1: decreased rate of (α2u) catabolism (renal proximal tubular cell lysosomes)

When a chemical binds to α2u, it forms a complex with the α2u protein that can result in a reduction in its rate of catabolism (Lehman-McKeeman et al. 1990). α2u then accumulates in the phagolysomes of the renal proximal tubules (Tables 1(A,B); Swenberg et al. 1989; Borghoff et al. 1990; US EPA 1991; IARC 1999; Swenberg and Lehman-McKeeman 1999). Although a decrease in the catabolism of this protein was directly measured in vitro in only one study (Lehman-McKeeman et al. 1990), this study demonstrated a response using isolated kidney cortical lysosomes with a number of the chemical stressors that cause α2u-N, or with their metabolite (e.g. d-limonene-1,2-oxide, isophorone, 2,5-dichlorophenol). Indirect measures of decreased catabolism were noted via studies that show an increase in α2u in the proximal tubule cells of exposed rats by immunohistochemical staining for α2u and or quantitation of α2u in kidney tissue (US EPA 1991, IARC 1999).

Key event 2: increased accumulation of α2u resulting in exacerbation of protein droplets (renal proximal tubular cells)

Under normal physiological conditions, α2u is degraded in the kidney by proteolytic enzymes within phagolysosomes. However, when the chemical initiator (chemical identified as binding to α2u) binds reversibly to α2u (MIE), it can decrease its rate of proteolytic hydrolysis, leading to its accumulation in the renal proximal tubular cells (summarized in Tables 1(A,B), see reviews by Swenberg et al. 1989; Borghoff et al. 1990; US EPA 1991; IARC 1999; Swenberg and Lehman-McKeeman 1999). Increased accumulation of α2u has been evaluated by quantitating α2u in the kidney using an enzyme-linked immunosorbent assay (ELISA), semi-quantitating by grading the size and extent of α2u accumulation in the kidney in the form of protein droplets histologically, and/or immunohistochemically staining the droplets for α2u. The protein droplets reflect the accumulation of exfoliated renal tubule cells, specifically in the P2 segment of the renal proximal tubules, where the majority of protein is observed. Under specific exposure concentrations including both acute and repeated exposure, certain chemical stressors (Tables 2 and 3; citations listed in Table 3) result in excessive protein droplet accumulation, described as crystal-like structures in the renal proximal tubules (i.e. 2,4,4-trimethyl pentane, d-limonene, 1,4-dichlorobenzene).

Table 3. Summary of examples from the evidence base that establish support for the dose-responses of KERs.

Chemical	Species	Route	Duration	Concentration/dose	Endpoint	Key event	References
MIBK	Rat	Inhalation	1 week	450*, 900, 1800 ppm	Increased protein droplets, droplets stained for α2u, and Increased α2u concentration	KE2	Borghoff et al. 2015
MIBK	Rat	Oral	4 weeks	450*, 900, 1800 ppm	Increased protein droplets, and Increased α2u concentration	KE2	Borghoff et al. 2015
MTBE	Rat	Oral	14 days	350, 700, 1000*, 1500* mg/kg/d	Protein droplet accumulation	KE2	Robinson et al. 1990
MTBE	Rat	Inhalation	4 week	400*, 3000*, 8000 ppm	Protein droplet accumulation	KE2	Bird et al. 1997
MTBE	Rat	Oral	90 days	100, 300, 900*, 1200* mg/kg	Cytoplasmic protein droplets in tubular epithelial cells	KE2	Robinson et al. 1990
MTBE	Rat	Inhalation	10 days	400*, 1500*, 3000 ppm	Protein droplet accumulation	KE2	Prescott-Mathews et al. 1997
TBA	Rat	Inhalation	10 days	250*, 450, 1750* ppm	Renal α2u concentration	KE2	Borghoff et al. 2001
TBA	Rat	Inhalation	13 weeks	230*, 490, 840, 1250 mg/kg	Granular casts	KE2	Hard et al. 2019
d-limonene	Rat	Oral	27 days	75*, 150, 300 mg/kg	Protein droplet accumulation	KE2	Kanerva et al. 1987
ETBE	Rat	Inhalation	1 week	500*, 1750, 5000 ppm	Protein droplet accumulation	KE2	Medinsky et al. 1999
ETBE	Rat	Inhalation	4 weeks	500**, 1750, 5000 ppm	Protein droplet accumulation	KE2	Medinsky et al. 1999
ETBE	Rat	Inhalation	13 weeks	500, 1750, 5000* ppm	Protein droplet accumulation	KE2	Medinsky et al. 1999
ETBE	Rat	Oral	4 weeks	300*, 1000 mg/kg	Protein droplet accumulation	KE2	Hagiwara et al. 2011
Decalin	Rat	Oral	3 months	25*, 50, 100, 400 mg/kg	Protein droplet accumulation	KE2	Doi et al. 2007
MIBK	Rat	Inhalation	1 week	450*, 900, 1800* ppm	Renal cell proliferation	KE3	Borghoff et al. 2015
MIBK	Rat	Oral	4 weeks	450*, 900, 1800 ppm	Renal cell proliferation	KE3	Borghoff et al. 2015
MIBK	Rat	Inhalation	70 days	500*, 2000 ppm	Regenerative tubular epithelia	KE3	Nemec et al. 2004
MTBE	Rat	Oral	14 days	350, 700, 1000*, 1500* mg/kg/d	Renal cell proliferation	KE3	Robinson et al. 1990
MTBE	Rat	Inhalation	10 days	400*, 1500*, 3000 ppm	Renal cell proliferation	KE3	Prescott-Mathews et al. 1997
MTBE	Rat	Oral	90 days	100, 300*, 900*, 1200 mg/kg	Renal cell proliferation	KE3	Robinson et al. 1990
MTBE	Rat	Inhalation	4 weeks	400*, 3000*, 8000 ppm	Renal cell proliferation	KE3	Bird et al. 1997
ETBE	Rat	Inhalation	1 week	500*, 1750, 5000 ppm	Renal cell proliferation	KE3	Medinsky et al. 1999
ETBE	Rat	Inhalation	4 weeks	500, 1750*, 5000* ppm	Renal cell proliferation	KE3	Medinsky et al. 1999
ETBE	Rat	Inhalation	13 weeks	0, 500, 1500, 5000* ppm	Renal cell proliferation	KE3	Medinsky et al. 1999
TBA	Rat	Inhalation	10 days	250*, 450, 1750 ppm	Renal cell proliferation	KE3	Borghoff et al. 2001
TBA	Rat	Inhalation	13 weeks	490*, 840 mg/kg	Renal cell proliferation	KE3	Hard et al. 2019
d-limonene	Rat	Oral	30 weeks	150* mg/kg	Renal cell proliferation	KE3	Dietrich and Swenberg 1991
d-limonene	Rat	Oral	13 weeks	150*, 300, 600, 1200, 2400 mg/kg	Renal regeneration	KE3	NTP 1990
Decalin	Rat	Oral	3 months	25*, 50, 100, 400 mg/kg	Renal tubular regeneration	KE3	Doi et al. 2007
1,4-dichlorobenzene	Rat	Oral	1, 4, 13 weeks	25*, 75, 150 300 mg/kg	Renal tubule proliferation	KE3	Lake et al. 1997
Isophorone	Rat	Oral	2 years	250*, 500 mg/kg	Atypical hyperplasia, linear mineralization	KE4	NTP 1986
d-limonene	Rat	Oral	2 years	75*, 150 mg/kg	Linear mineralization	AE	NTP 1990
Decalin	Rat	Inhalation	2 years	25*, 50, 100 400 ppm	Linear mineralization	AE	NTP 2005
ETBE	Rat	Oral	4 weeks	300*, 1000* mg/kg	Atypical hyperplasia	AE	Hagiwara et al. 2011
ETBE	Rat	Inhalation	2 years	500, 1500*, 5000* ppm	Mineral deposition	AE	Saito et al. 2013
MIBK	Rat	Inhalation	2 years	450*, 900, 1800 ppm	Linear mineralization	AE	NTP 2007
TBA	Rat	Inhalation	2 years	90, 200*, 400*mg/kg	Linear mineralization	AE	Hard et al. 2019
Decalin	Rat	Inhalation	2 years	25, 50*, 100* ppm	Renal tubule adenoma/carcinoma (combined)	AO	NTP 2005
d-Limonene	Rat	Oral	2 years	75*, 150 ppm	Renal tubule adenoma/carcinoma (combined)	AO	NTP 1990
Isophorone	Rat	Oral	2 years	250*, 500 ppm	Renal tubule adenoma/carcinoma (combined)	AO	NTP 1986
MIBK	Rat	Inhalation	2 years	450*, 900*, 1800 ppm	Renal tubule adenoma/carcinoma (combined)	AO	NTP 2007
MTBE	Rat	Inhalation	2 years	400*, 3000*, 8000 ppm	Renal tubule adenoma/carcinoma (combined)	AO	Bird et al. 1997
TBA	Rat	Inhalation	2 years	90*, 200, 420 mg/kg	Renal tubule adenoma/carcinoma (combined)	AO	Hard et al. 2019

Bolded text indicates significant changes in dose-groups reported for endpoint measured; * indicates that the data point is shown in Figure 1 as either the highest no effect level or the lowest effect level (LEL) reported per reference; ** indicates that only one ‘medium term’ LEL is shown for this study although this LEL was observed at both 28 and 90 days in the reference. Note that each key event is not represented for every chemical in Figure 1; only key events with observations for a chemical are included in Figure 1.

Protein droplet accumulation was also reported in male rats following 14 and 90 days of oral exposure to MTBE (Robinson et al. 1990). Prescott-Mathews et al. (1997) demonstrated an increase in protein droplet accumulation with droplet staining for α2u and an increase in the concentration of α2u in the kidney of male rats following 10 days of exposure to varying concentrations of MTBE. A time- and concentration-dependent increase in protein-droplet accumulation was measured in male rat kidneys, with evidence of protein droplet accumulation following inhalation to ETBE for 90 days (Medinsky et al. 1999). Alpha-2u immunoreactivity in these protein droplets was confirmed. Another study showed concentration-dependent accumulation of protein droplets in the kidneys of male rats, following 10-day inhalation to TBA (0, 230, 450, and 1750 ppm) (Borghoff et al. 2001). In a more recent study, rats were exposed via inhalation for 1 or 4 weeks to 0, 450, 900, or 1800 ppm MIBK for 6 h/day. An increase in protein droplet formation (also referred to as hyaline droplets in H&E stained kidney sections) was reported to occur following exposure to 450 ppm and higher, increasing in incidence and severity with increasing concentration (Borghoff et al. 2015). This KE is consistently reported to occur in multiple studies following in vivo exposure to different α2u chemical stressors across various exposure scenarios and duration (Phillips et al. 1987; Robinson et al. 1990; Medinsky et al. 1999; Prescott-Matthews 1999; Borghoff et al. 2001, 2009, 2015; Nemec et al. 2004). Several review articles and regulatory assessments also identify chemicals that cause this key event (see reviews by US EPA 1991; Swenberg 1993; IARC 1999; Swenberg and Lehman-McKeeman 1999).

Key event 3: cytotoxicity-induced cell proliferation (renal proximal tubular cells)

The accumulation of α2u (KE2) within the phagolysosome results in cytotoxicity of the kidney tubular cells (Tables 1(A,B); see Borghoff et al. 1990; Hard et al. 1993; Swenberg and Lehman-McKeeman 1999; Borghoff et al. 2001). Male rat α2u-specific lesions result from recurring epithelial cell death, and ensuing cell shedding into the lumen within the renal proximal tubules, and result in the formation of granular casts at the change-over from the pars recta of the proximal tubules and the descending Thin-Loop of Henle due to cellular debris overload and subsequently the accumulation of cell mineralization products in the collecting duct and ductus of Bellini of the papilla (Hard et al. 2013; Borghoff et al. 2015). Compensatory cell proliferation takes place in an effort to overcome recurring cell death and is found to be specific to the site at which α2u accumulation occurs (Swenberg and Lehman-McKeeman 1999; Borghoff et al. 2001).

Many of the chemical stressors identified in Tables 2 and 3 as operating through this MoA to cause kidney tumors in male rats demonstrated an increased and sustained cell proliferative response. Increased kidney cell proliferation has been shown to occur following TMP and DCB exposure (Eldridge et al. 1990; Lake et al. 1997) and shown to increase with dose and time in male rats only, following inhalation exposure for 90 days to ETBE (Medinsky et al. 1999), and up to 10 days following exposure to MTBE or TBA (Prescott-Mathews et al. 1997; Williams and Borghoff 2001). TBA was also reported to have an increased cell proliferation in the male rat kidneys, corresponding to increased concentrations of α2u (Williams and Borghoff 2001).

Key event 4: atypical hyperplasia (renal tubules)

Sustained regenerative cell proliferation described in KE3 leads to atypical hyperplasia in the presence of renal lesions characteristic of protein accumulation, such as granular cast formation and linear mineralization (Tables 1(A,B); see NTP 1986, 2007; Borghoff et al. 2001, Doi et al. 2007; Saito et al. 2013). Multiple studies report atypical hyperplasia following exposure to different chemicals that also have reported previous KEs in this AOP, as summarized in Figure 1 and Table 3. Atypical renal tubular hyperplasia is described by Doi et al. (2007) to consist of discrete, focal to multifocal proliferative lesions of renal tubules. It is noted that grading of hyperplasia is based on the size of the lesion, the number of affected tubular profiles present, and the complexity of the hyperplastic epithelium. Granulation cast formulation and linear mineralization are considered an associative event (AE) that acts as a reliable indicator of atypical hyperplasia but is not considered a biological process in and of itself that is necessary to progress through the overall pathway (Andersen et al. 2014; Simon et al. 2014). Linear mineralization is described as regularly arranged basophilic mineralized linear deposits that are distributed mainly in the lower half of the renal papillae. d-Limonene induced atypical renal tubule hyperplasia in male rats, but not females, as reviewed in Rice et al. (1999). NTP (2007) reports atypical renal tubule hyperplasia in male rats following inhalation exposure to MIBK for 2 years; Hagiwara et al. (2011) report similar findings in male rats following oral exposure to ETBE for 4 weeks. Sustained regenerative cell proliferation leads to atypical hyperplasia, which is believed to be associated with the promotion of spontaneous mutations through clonal expansion.

Adverse outcome (AO): kidney adenomas/carcinomas

Sustained cell proliferation and clonal expansion of the renal tubular cells lead to a cellular environment leading to tumorigenic response (Tables 1(A,B); see Borghoff et al. 1990, 2001, 2015; Borghoff and Lagarde 1993; Lehman-McKeeman et al. 1990; Swenberg 1993; Prescott-Matthews et al. 1999). The extent of this sustained cell proliferation is seen in the observation of atypical hyperplasia in the kidneys of male rats. Inhalation exposure to 3000 ppm MTBE caused an increased incidence of kidney tumors in male rats, but not female rats or mice (Bird et al. 1997; Prescott-Matthews et al. 1999). TBA was also reported to induce kidney tumors in male rats following exposure in drinking water (Cirvello et al. 1995; NTP 1995). A diverse group of chemicals has been identified to cause kidney tumors in male but not female rats (Swenberg et al. 1989; Borghoff et al. 1990; US EPA 1991; IARC 1999; Swenberg and Lehman-McKeeman 1999; Doi et al. 2007).

AOP assessment

Defining the chemical and biological domains of applicability

The MIE applies to several chemicals across diverse classes of chemicals, all of which bind reversibly as the parent chemical or a metabolite, to the α2u protein. Empirical evidence demonstrating activation of MIE and KEs was reviewed for 1,4-dichlorobenzene (2,5-dichlorophenol), isophorone, d-limonene (d-limonene 2-5-oxide), decalin, MIBK, ETBE (TBA), MTBE (TBA), TBA, and TMP (TMPOH) (Table 2).

This AOP biological domain of applicability is specific to sexually mature adult male rats (∼10–12 weeks, peak of α2u synthesis) which would be also the relevant life stage. The hepatic synthesis of α2u in male rats begins at the onset of sexual maturity and increases until ∼20 weeks of age (Vandoren et al. 1983; Roy et al. 1983). α2u is synthesized at a high level in the liver of male but not female rats (Roy et al. 1976; Burnett et al. 1989), although there are glands in female rats that synthesize α2u, this protein is only present at extremely low levels. α2u kidney concentrations are ∼120 times lower in female kidneys vs. male rat kidneys, blood, and urine (Vandoren et al. 1983; Williams and Borghoff 2001). Chemical binding to α2u and subsequent nephropathy has consistently been reported to occur only in male, but not female, rats (Alden 1986; Swenberg et al. 1989; Borghoff et al. 1990; Ridder et al. 1990; Dietrich and Swenberg 1991; Swenberg 1993; Doi et al. 2007; Aksu and Tanrikulu 2016).

α2u has not been detected in human kidneys; one study showed that a similar protein to α2u, could not be detected in human kidneys based on its abundance and chemical binding specificity (Borghoff and Lagarde 1993). In addition, mice excrete a mouse urinary protein similar to α2u, but its amino acid sequence homology is ∼90% that of rat α2u (Shaw et al. 1983) and ligand specificity and reabsorption in the kidney differs from rat α2u. Female rats, NBR male rats, or mice, which do not synthesize α2u in the liver, do not develop α2u-mediated nephropathy or kidney tumors when exposed to the chemicals that cause these responses in the male rats (Alden 1986; Short, Burnett, et al. 1989; Short, Steinhagen, et al. 1989; Swenberg et al. 1989; Borghoff et al. 1990; Ridder et al. 1990; Dietrich and Swenberg 1991, 1992a, 1992b; US EPA 1991; Swenberg 1993; Doi et al. 2007; Aksu and Tanrikulu 2016). Again supporting that this AOP’s taxonomic applicability is specific to the male rat (Rattus norvegicus).

Assessing the relative confidence in the AOP

The justification and support for the essentiality of the KEs based on biological plausibility (Table 4), essentiality (Table 5), and empirical evidence (Table 6) are therefore provided below with detailed tables and summaries that are consistent with these OECD templates for assessment of confidence of an AOP. The confidence in KEs and KE relationships is summarized in Table 1(B).

Table 4. Summary assessment of the relative level of confidence in the overall AOP based on weight-of-evidence elements—support for the biological plausibility of KERs.

		Defining question	High (strong)	Moderate	Low (weak)
Support for biological plausibility of KERs	Direct or indirect KER	Is there a mechanistic relationship between KEup and KEdown consistent with established biological knowledge?	Extensive understanding of the KER based on extensive previous documentation and broad acceptance	KER is plausible based on analogy to accepted biological relationships, but scientific understanding is incomplete	Empirical support for association between KEs, but the structural or functional relationship between them is not understood
KER1	Direct	Strong: The evidence to support the direct relationship between the MIE (chemical binding to α2u) and KE1 (inhibition of α2u catabolism) is supported by data reported for diverse chemical structures within the context of a well-defined mode of action, based on empirical evidence both in vitro and in vivo. Therefore, it is widely accepted that chemical binding to α2u directly affects the catabolism of α2u.
KER2	Indirect	Strong: Available data for multiple chemicals support a finding that the reversible binding of the chemical to α2u results in a chemical-protein complex that makes the protein more resistant to catabolism. Data are available that demonstrate an increase in and reversibility of the protein droplet in vivo that slows the rate of catabolism.
KER3	Direct	Strong: Consistent evidence is reported for multiple chemicals on the formation of protein droplets—immunohistochemical staining of α2u within the droplet, and or quantitation α2u concentration in kidneys.
KER4	Direct	Strong: It is well-supported that there is a direct correlation between α2u protein droplet accumulation, cytotoxicity, and regenerative cell proliferation following exposure to one of several different chemical stressors (see Stressor Table 2)
KER5	Direct	Strong: There is evidence for sustained cell proliferation associated with the development of atypical hyperplasia, along with these events, which have the potential to result in pre-neoplastic and neoplastic lesions in the kidney.

Table 5. Support for the essentiality of KEs.

Support for essentiality of Kes	Defining question	High (strong)	Moderate	Low (weak)
	Are downstream KEs and/or the AO prevented if an upstream KE is blocked?	Direct evidence from specifically designed experimental studies illustrating essentiality for at least one of the important KEs	Indirect evidence that sufficient modification of an expected modulating factor attenuates or augments a KE	No or contradictory experimental evidence of the essentiality of any of the KEs
MIE	There is evidence for chemical binding to α2u and that this binding is reversible based on both in vivo and in vitro evidence. When chemical exposure is removed, protein droplets (or concentrations of α2u) are resolved with time as the complex is catabolized. This is demonstrated in time course experiments in which there is no evidence of protein accumulation a few days after chemical administration is stopped. Along with a demonstration of chemical binding when α2u is isolated from male rat kidney with along with the measured thereby supporting the essentiality of this MIE and downstream events for this AOP.
KE1	There is direct evidence that chemical binding to α2u decreases its catabolism and the complex results in an increase in the protein in the proximal tubules in the form of droplets. Once chemical administration is stopped, the protein droplets that accumulate are resolved. The lack of α2u in female rats and other strains/species, with the lack of these chemicals inducing α2u-N (i.e. protein droplet accumulation) supports the essentiality of this KE.
KE2	Direct evidence for the accumulation of α2u in the proximal tubules was reported in several studies either through protein droplet assessment histologically, staining of the droplets for α2u immunohistochemically, or quantitating the α2u concentration. Together with the lack of α2u in female rats and other strains/species, the lack of these chemicals inducing α2u-N (i.e. protein droplet accumulation) supports the essentiality of this KE.
KE3	While there is no direct experimental evidence on the recovery or reversibility of the KE and subsequent key events, there are data from other species and sex—specifically, female rat and mice data that support the establishment of this KE. The lack of data in other species and female rats supports the essentiality of this KE.
KE4	While there is no direct experimental evidence on the reversibility of this KE, the lack of α2u in female rats and other strains/species, with the lack of these chemicals inducing atypical hyperplasia and or linear mineralization, supports the essentiality of this KE.
AO	The lack of α2u in female rats and other strains/species, with the lack of these chemicals inducing KE2-KE4, supports the essentiality of this AO.

Table 6. Empirical evidence for KERs.

Empirical evidence for KERs	Defining question	High (strong)	Moderate	Low (weak)
	Does empirical evidence support that a change in KEup leads to an appropriate change in KEdown? Does KEup occur at lower doses and earlier time points than KE down, and is the incidence of KEup > than that for KE down? Inconsistencies?	Multiple studies show a dependent change in both events following exposure to a wide range of specific stressors. No or few critical data gaps or conflicting data.	Demonstrated dependent change in both events following exposure to a small number of stressors. Some inconsistencies with an expected pattern can be explained by various factors.	Limited or no studies reporting a dependent change in both events following exposure to a specific stressor, and/or significant inconsistencies in empirical support across taxa and species that don’t align with hypothesized AOP
KER1	Strong: There is strong direct evidence that chemical binding to α2u inhibits its catabolism based on an in vitro study conducted with several stressors or their metabolites. There is strong indirect evidence that chemical binding to α2u is associated with decreased catabolism in vivo, based on protein accumulation, and when α2u is isolated from the male rat kidney, the chemical or metabolite co-elutes with this protein. Because liver α2u synthesis nor increased reabsorption has been shown to change with exposure to a chemical stressor, the body of evidence is strong for demonstrating this relationship.
KER2	Strong: In studies conducted with a number of the identified chemical stressors, clear evidence for temporal, dose-response, and incidence concordance for protein droplet exacerbation in male but not female rats has been demonstrated.
KER3	Strong: Several studies using multiple types of stressors have shown a direct relationship, and temporal, dose-response, and incidence concordance with the formation of protein droplet formation, leading to cytotoxicity and compensatory cell proliferation in the renal tissue.
KER4	Moderate: Although there is clear and consistent evidence for the relationship between cytotoxicity and compensatory cell proliferation, there is only an association of these measures with atypical hyperplasia, based mainly on the difficulty in quantitating these lesions. However, with increased cell proliferation that is sustained in the male rat kidney and associated with mainly α2u-N measures, atypical hyperplasia along with linear mineralization have only been identified in the male but not female rats.
AO	Moderate: There is moderate empirical evidence that shows concordance of consistency, temporal occurrence, and dose-response that is not demonstrated with many stressors based on low renal tumor incidence, however, there is incidence concordance of hyperplasia and the tumorigenic response. This is also a well-accepted outcome associated to occur with the general mechanism of cancer.

WOE: biological plausibility

The biological plausibility of each KER within this pathway is considered to be strong (high, see Table 4), based on the well-understood mechanistic basis of the KERs that are well-established, documented, and widely accepted in the scientific community. Briefly, the evidence to support the direct relationship between the MIE (chemical binding to α2u) and KE1 (reduction of α2u catabolism) (KER1) is supported by data reported for diverse chemical structures within the context of a well-defined MoA and based on empirical evidence in vitro and in vivo. Indirect data from multiple chemicals in vivo are available that demonstrate the reversible binding of the chemical to α2u (KER2) resulting in a chemical-protein complex. There is also consistent direct evidence in vivo for multiple chemicals for the formation of protein droplets (KER3), observed through the immunohistochemical staining of α2u within the droplet, and quantitation of α2u concentration in kidneys. It is also well-established and widely accepted there is a direct correlation between protein droplet accumulation, cytotoxicity, and regenerative cell proliferation (KER4); which then through sustained cell proliferation leads to the development of atypical hyperplasia.

WOE: essentiality

The essentiality of the early KEs in this pathway is considered strong (MIE, KE1, KE2), based on direct experimental evidence showing that elimination of one of the upstream KEs prevents progression through this pathway (Table 5). Later KEs (KE3, KE4, and AO) are considered moderately essential to this AOP as there was no direct experimental evidence available that illustrated the essentiality of these KEs. However, the lack of hepatic synthesis of α2u in female rats, and other species including humans, with the lack of the identified chemical stressors inducing KE2-KE4 and the direct empirical evidence for earlier KEs supports the essentiality of these KEs and the overall AOP.

WOE: empirical evidence

Individually, relationships between early KEs are supported by strong empirical evidence, based on available direct and indirect evidence across multiple chemicals demonstrating that binding to α2u is associated with decreased catabolism and protein droplet accumulation, eventually leading to cytotoxicity and compensatory cell proliferation in the kidney tissue (i.e. KERs between MIE, KE1, KE2, KE3) (Table 6). Relationships between the later KEs are considered to be supported by moderate evidence, mainly due to the difficulty quantitating atypical hyperplasia and the reported low incidence of kidney tumors in the available studies (i.e. relationship between KE3, KE4, and AO).

Collectively, this AOP was developed based on empirical evidence for 9 distinct chemicals with generally consistent dose-temporal patterns (Table 2). For example, acute (short) and sub-chronic (medium) exposure to ETBE at doses of ∼300–500 ppm induces KE2 but not KE3 and KE4 (Medinsky et al. 1999; Hagiwara et al. 2011). Induction of KE3 and KE4 is observed after relatively higher or relatively longer doses of ETBE (≥1000 ppm) (Medinsky et al. 1999; Hagiwara et al. 2011). Although this pattern of dose-responsiveness across KEs is also observed for MTBE, there is a noted inconsistency for TBA, where KE2 is not observed after acute exposure of 250 ppm, but KE3 occurs at this dose (Borghoff 2001). Given the integration of studies conducted under varying methods (e.g. test material, route of exposure, study duration, species), the dose-response supports moderate-to-strong confidence in this AOP.

WOE: limitations and/or inconsistencies

As noted above, not all KERs are supported by strong biological plausibility and empirical evidence. Further, a dose-response relationship has not been demonstrated directly in the MIE or KE1. However, Borghoff et al. (1991) demonstrated different binding affinities of the chemical stressors to α2u in vitro, but a direct relationship between the amount of protein accumulation and binding affinity is difficult to measure in vivo, along with the complexity associated with either or both the parent and metabolite binding to the protein, and not necessarily both affecting catabolism (Lehman-McKeeman et al. 1990).

Discussion

The purpose of this assessment was to elucidate and assess the confidence in an AOP for male-rat-specific kidney toxicity of α2u-N induced kidney tumors through the evaluation of primary literature, reviews, and regulatory guidance.

This assessment developed a high confidence AOP for kidney tumors in male rats based on an MIE of reversible binding of chemical stressors to the α2u-globulin protein. This chemical binding leads to a reduction in the catabolism rate of α2u (KE1); resulting in the accumulation of the α2u in the form of protein droplets in the lysosomes of RPT cells (KE2). Exacerbation of α2u accumulation in the RPT cells results in cytotoxicity and compensatory cell proliferation (KE3). Chronic cytotoxicity and regenerative cell proliferation leads to atypical hyperplasia (KE4) that is indicated by mineralization of the RPTs (AE) leading to, with continued chronic exposure, the formation of kidney adenomas and carcinomas in male rats (AO).

Collectively, there is moderate-to-strong support in the WOE for the overall assessment of this AOP based on the essentiality of the KEs, biological plausibility, and empirical evidence of the KERs. Considering the empirical evidence, some inconsistencies are noted. Inhalation exposure of MIBK for 1 week, but not 4 weeks activated KE3 at lower but not higher doses (450 ppm but not 900 ppm) (Borghoff 2015) and inhalation exposure to TBA (250, 450, or 1750 ppm) for 10 days activated the downstream (KE3) but not the upstream KEs (KE2) at 250 ppm (Borghoff et al. 2001). However, inconsistencies, such as these are best described as limited when compared to the totality of data across other chemical stressors and are thought to be influenced by the sensitivity of the assay and the overall mild induction of α2u-N and sensitivity of the assays that measure these early events. The overall WOE still ranges in strength from moderate to strong. Consequently, the AOP as a whole is considered to have high confidence.

Based on the high overall confidence, the AOP for α2u-N MoA for kidney tumors could be used to evaluate the likelihood that chemicals act through this pathway or to determine the human relevance of kidney findings in animal studies. With an available AOP of high confidence, this tool can be used to develop new approaches to screen for this particular outcome after chemical exposure using biological events that occur earlier than the adverse outcome itself Characterization of the MIE as an indicator for the AO supports the creation of rapid screening of new and emerging chemicals through in vitro assays that can be used to identify more targeted testing strategies for hazard characterization. A similar approach has been applied to developmental neurotoxicity, where a network of AOPs has been developed with the intent to progress the development of in vitro and in silico testing approaches for this outcome (Bal-Price et al. 2015). Given the complexity of the developing brain, a series of AOPs were developed, with a focus on the identification of common KEs to reduce the overall resource investment in testing while increasing the likelihood that developmental perturbations are identified. The development of in vitro test batteries based on this common KE is envisioned to be used to prioritize higher-tier testing in animals in a data-informed, resource-conscious manner.

In the case of kidney outcomes, the AOP described here is envisioned to eventually be one in a series of kidney-focused AOPs, that could be used to confirm male rat-specific observations without progressing to chronic exposure studies. Although this pathway has long been recognized, it remains relevant and timely for regulatory risk assessment activities, particularly for hazard identification, prioritization of chemicals of potential concern for human health, product stewardship, and regulatory submissions.

A second commonly observed and interfering kidney lesion is chronic progressive nephropathy (CPN), a progressive rat kidney pathology that often co-occurs with α2u-N. CPN is considered a risk factor for the development of atypical tubule hyperplasia and renal tubule tumors. The hallmark features of CPN are histopathological events with progressive severity. It has been suggested that the likely cause underlying the association of CPN with renal tubule neoplasia is CPN-associated renal tubule cell proliferation (Hard et al. 2013). Distinct from α2u-N, scientific consensus has not been reached on whether the biological events leading to CPN are indeed specific to rodents, and thus the co-occurrence of α2u-N and CPN can be considered to be potentially human-relevant, as in the recent case of ETBE and TBA (US EPA 2021). An area of future work is to develop a comparative set of KEs that can be used to inform the outstanding questions of whether the biological events leading to CPN qualitatively occur in cells derived from human tissues.

The availability of a network of AOPs leading to kidney outcomes, specifically for both α2u- and CPN-mediated effects, might have helped clarify this question of human relevance earlier in the IRIS review process by providing a transparent and scientifically validated alternative to characterize the relative likelihood that the kidney effects in question were mediated via one of these pathways.

To that end, our analysis represents the first fully elucidated and empirically supported AOP for α2u-N-mediated kidney tumors in male rats established according to OECD guidance (OECD 2018). Earlier work with a similar concept is stated to be under development in a public database (AOP-Wiki), entitled “Alpha2u-microglobulin cytotoxicity leading to kidney tubular adenomas and carcinomas (in male rat)” and listing a similar order of key events and biological applicability domain. However, no detailed characterization is available (Wood 2020). The KEs identified by Wood (2020) and in this manuscript are well-established in the literature and have a broad acceptance in the scientific community and by both EPA and IARC (US EPA 1991; Swenberg and Lehman-McKeeman 1999). The novelty of the work herein is the integration of mechanistic evidence from a variety of sources and methodologies into the AOP with an explicit statement of confidence in the WOE. This work takes into consideration biological plausibility, essentiality, and empirical evidence for the KEs and also characterizes the dose-response for several chemicals.

In the future, this AOP could be applied to risk assessment in a few different ways. Firstly, it could be used to clarify the mechanistic basis of kidney effects observed in animal studies to assess the potential for human relevance. Building on the ETBE example above, in vitro assays could be developed and used to investigate ETBE’s ability to activate early KEs in the AOP: in vitro binding affinity assays to evaluate the interaction between the chemical and α2u (as described in Borghoff et al. 1991), measurements to quantify the reduction of lysosomal degradation of chemically-bound α2u (as described in Lehman-McKeeman et al. 1990), or possibly ex vivo kidney slices or 3 D microphysiological systems of the proximal tubule to assess subcellular localization of α2u in phagolysosomes (Saitta et al. 2019; Yeung and Himmelfarb 2019). Using, such assays may help to reduce the number of animals used to confirm if a chemical is likely to cause kidney tumors in male rats through the α2u-N MoA vs. conducting chronic in vivo animal studies. Kidney tissues from different rodent species, strains, and sexes could be assessed, including human-derived tissues or cell lines, to inform the species-specificity of observed changes in these in vitro assays. Secondly, the availability of a defined AOP with specified in vitro assays for alternate kidney mechanisms would allow research programs to make comparative assessments of the likelihood that each mechanism contributes to apical kidney outcomes. Drawing from existing animal studies on a diverse set of chemistries to develop this kidney-specific AOP is a step toward a paradigm shift in regulatory toxicology, where data-driven testing strategies may be developed based on screening MIEs drawn from a library of AOPs rather than by conducting a standard set of resource-intensive toxicology studies. In this way, the establishment of a library of high-confidence AOPs can facilitate achieving regulatory targets set out by organizations, such as the US EPA (US EPA 2019).

In conclusion, this work established a high-confidence AOP for male rat kidney tumors through the accumulation of the male rat-specific protein α2u. This AOP will be useful both for chemical screening using NAMs developed for measuring the MIE with critical KEs (e.g. decreased α2u catabolism) for use in either prioritization of chemical testing or hazard characterization in regulatory settings.

Abbreviations
α2u	=	α2u-globulin (α2u)
α2u-N	=	α2u nephropathy
AE	=	associative event
AO	=	adverse outcome
AOP	=	adverse outcome pathway
CPN	=	chronic progressive nephropathy
Da	=	daltons
DCB	=	dichlorobenzene
ELISA	=	enzyme-linked immunosorbent assay
ETBE	=	ethyl tertiary butyl ether
IARC	=	International Agency for Research on Cancer
KE	=	key event
KER	=	key event relationships
MIBK	=	methyl isobutyl ketone
MIE	=	molecular initiating event
MoA	=	mode of action
MTBE	=	methyl tert-butyl ether
NAM	=	new approach methods
NBR	=	NCI-Black-Reiter
OECD	=	Organization for Economic Coordination and Development
RPT	=	Renal proximal tubule
TBA	=	tertiary butyl alcohol
TMP	=	2,2,4-Trimethylpentane
TMPOH	=	2,4,4-Trimethylpentane-2,4,4-trimethyl-2-pentanol
U.S. EPA	=	United States Environmental Protection Agency
WOE	=	weight of the evidence

Acknowledgments

The authors wish to gratefully acknowledge API’s Toxicology Methods & Research Group for their encouragement to write this manuscript and for subsequent review of this manuscript, specifically Pearl Moy (Chevron and member) who provided helpful editorial comments; API’s Health and Product Stewardship Sub-Committee for financial support and review of the manuscript; and to the external reviewers selected by the Editor and anonymous to the authors whose comments were extremely valuable in revising and refining the manuscript.

Declaration of interest

KG is an employee of ExxonMobil Biosciences, Inc., a separately incorporated but wholly owned affiliate of Exxon Mobil Corporation. SS is an employee of Shell International and is free to express his own scientific opinion and ideas. Both ExxonMobil and Shell manufacture gasoline and both are members of the American Petroleum Institute (API), a trade association that represents all segments of oil and gas development, including manufacturers of gasoline. ETBE and TBA (chemicals reviewed in this manuscript) are fuel oxygenates that have historically been added to gasoline and have recently been reviewed by the US EPA’s IRIS Program. JR was an employee of API during the development of this manuscript but is currently employed elsewhere. At API, JR managed both the API Toxicology Methods and Research Group, on which KG participates, and the API Health and Product Stewardship Sub-Committee, on which SS acts as the chair. This Group and Sub-Committee commissioned and funded this work, as well as reviewed and approved the manuscript. AF and SB are employees of ToxStrategies, Inc., a scientific consulting firm. ToxStrategies, Inc. was retained as paid consultants to produce the API-funded report. An early version of this work was presented by AF and SB as an e-poster at the Society of Toxicology 59th Annual Meeting (#1997), which was also approved by the API Group before the presentation. The work of all Authors on this manuscript was done as part of their normal employment. Before submission, this manuscript underwent internal, legal review. This internal review did not impact the text, which was written by the Authors. Some Authors have participated in regulatory proceedings on chemicals cited in this manuscript: KG and JR provided public comments to the US EPA on ETBE and TBA IRIS assessments on behalf of API and SB provided public comments on behalf of Lyondell Chemical Company; SS/SB participated in trade association panels sponsoring research to inform IARC assessments, and SB participated in IARC reviews for several of these chemicals. In addition to regulatory proceedings, readers should be aware that chemicals mentioned in this publication that support the proposed AOP have been, currently are, or possibly will be, involved in civil litigation. This civil ligation may involve API member companies.

Data availability statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.