The cancer mode of action for 1,4-DX plays a critical role in estimating cancer potency and setting threshold levels for this chemical. Please indicate your confidence rating on the weight of evidence supporting/refuting the potential mode of action for the carcinogenic effects of 1,4-DX in the rodent liver (-5=strong refuting evidence, 0=lack of or equivocal evidence, 5=strong supporting evidence)

Answer Explanations

• View answer

  Expert 5

  	-5	-4	-3	-2	-1	0	1	2	3	4	5
  Direct genotoxicity	1	0	0	0	0	0	0	0	0	0	0
  Indirect genotoxicity	1	0	0	0	0	0	0	0	0	0	0
  Proliferative regeneration induced by cytotoxicity	0	0	0	0	0	0	0	0	0	1	0
  Oxidative stress	0	0	0	0	0	1	0	0	0	0	0
  Activation of nuclear receptors and associated transcription factors (e.g., CAR/PXR)	0	0	0	0	0	0	1	0	0	0	0
  Other (please explain)

  Oxidative stress
  The issue of oxidative stress induction as a mode of action  is compounded by the fact that toxicity (necrosis) and resulting inflammation results in oxdiative stress and damge. Oxidative damage and stress should be considered a part of the cytotoxicity pathway.  Oxidative damage could also modify the response seen in the surviving (non necrotic) proliferating cells , thus enhancing the response.   I am not convinced that oxidative stress /damage is a key event.  Even in the case of metal overload (Fe and Cu) in the liver which produces oxidatve stress/damage, the liver tumor induction appears to be through cytotoxicity.   
  CAR/PXR
  The metabolism of 1,4 DX appears to be through CYP2b and/or 2E.  Therfore the participation of CAR in the CYP 2b pathway is expected.  While many of the key events linked to a cytotoxic MOA overlap with the CAR (another nuclear receptor MOA), this is expected since the multistep liver cancer process is similar subsequent to the initial key event.  While the data are not strong for a CAR MOA a further review of the availble results and possibly an additional experiemnt specifically address CAR activation might be warranted. 
  
  
• View answer

  Expert 6

  	-5	-4	-3	-2	-1	0	1	2	3	4	5
  Direct genotoxicity	1	0	0	0	0	0	0	0	0	0	0
  Indirect genotoxicity	0	0	0	0	0	0	0	0	0	1	0
  Proliferative regeneration induced by cytotoxicity	0	0	0	0	0	0	0	0	0	0	1
  Oxidative stress	0	0	0	0	0	0	0	1	0	0	0
  Activation of nuclear receptors and associated transcription factors (e.g., CAR/PXR)	0	0	0	0	0	0	1	0	0	0	0
  Other (please explain)	0	0	0	0	0	0	0	1	0	0	0

  Mo et al. (2022) show data suggesting an activation of TRPV1, vanilloid receptor by 1,4-DX. This may open a link to calcium influx, hypertrophy, necrosis/apoptosis, cell death. One alternative option may be 'oxidative stress'??
• View answer

  Expert 4

  	-5	-4	-3	-2	-1	0	1	2	3	4	5
  Direct genotoxicity	0	0	1	0	0	0	0	0	0	0	0
  Indirect genotoxicity	0	0	0	0	0	0	0	0	0	1	0
  Proliferative regeneration induced by cytotoxicity	0	0	0	0	0	0	0	0	1	0	0
  Oxidative stress	0	0	0	0	0	0	0	1	0	0	0
  Activation of nuclear receptors and associated transcription factors (e.g., CAR/PXR)	0	0	0	0	1	0	0	0	0	0	0
  Other (please explain)	0	0	0	0	0	0	0	0	1	0	0

  Based on current data and particularly, recent studies relevant to consideration of early key events (e.g., Chen et al., 2022; Lafranconi et al., 2023), the most likely mode of action for the carcinogenic effects of 1,4-DX in the rodent liver appears to be indirect genotoxicity, mediated through the induction of cyp2E1 metabolism, oxidative stress, DNA damage, hepatotoxicity, sustained proliferation and tumours.  This MOA is consistent with key events described in two existing AOPs, namely OECD 220 (https://aopwiki.org/aops/2200) and Cho et al., 2022 (DOI: 10.1002/em.22479).  This hypothesized MOA subsumes several of the options noted in this question (oxidative stress and proliferative regeneration) as key events; the confidence ratings should be considered in this context. (i.e., confidence in an indirect genotoxic mode of action involving early oxidative stress and later proliferative regeneration, the basis for the response to "other"). Responses for oxidative stress and proliferative regeneration are based on their likelihood of being key events in the hypothesized mode of action.
  
  With respect to the distinction of "direct" and "indirect" genotoxicity, U.S. EPA (2007) indicates that: “Key data for a mutagenic mode of action may be evidence that the carcinogen or a metabolite is DNA reactive and/or has the ability to bind DNA. Also, such carcinogens usually produce positive effects in multiple test systems for different genetic endpoints, particularly gene mutations and structural chromosome aberrations, and in tests performed in vivo which generally are supported by positive tests in vitro.” (https://archive.epa.gov/osa/mmoaframework/web/html/index.html). There is no evidence that 1,4-DX is reactive with DNA nor that it is metabolized to reactive electrophiles. Based on a relatively robust dataset in both in vitro and in vivo assays, 1,2-DX is not mutagenic and only weakly genotoxic (US EPA, 2013). The pattern of positive results in genotoxicity assays (namely DNA damage and single strand breaks, cell transformation and sister chromatid exchange in vitro and DNA damage, micronucleus formation and DNA synthesis in vivo generally at high concentrations) doesn't necessarily imply direct interaction (reactivity) with DNA; for example, the positive results for replicative DNA synthesis provide support for a non-genotoxic mode of action. The nature of reported DNA adducts following 1,4-DX exposure (i.e., 8-oxo-dG, a well-known oxidative damage biomarker but no DNA alkylation) is also supportive of an indirect mode of action involving oxidative stress. The additional observation by US EPA (2013) that responses in genotoxicity assays were positive generally only “in the presence of toxicity” is also consistent with genetic toxicity being a secondary consequence of earlier key events such as oxidative stress and/or hepatic damage/persistent regenerative proliferation.
  
  The extent of the evidence for hypothesized modes of action is normally considered through formal analysis based on the modified Bradford Hill considerations of biological plausibility, essentiality of key events, empirical support (dose-response and temporal concordance), consistency and analogy [US EPA, 2005 (https://www.epa.gov/risk/guidelines-carcinogen-risk-assessment), WHO  IPCS,2014,(https://www.who.int/publications/i/item/9789241563499), and OECD, 2023, (https://aopwiki.org/handbooks/4)]. These considerations are rank ordered as listed above with biological plausibility being most influential, followed by the essentiality of key events and so on.  
   

  The biological plausibility of the MOA hypothesized above is well documented in existing AOPs (https://aopwiki.org/aops/2200 and Cho et al., 2022 (DOI: 10.1002/em.22479) though the availability of data on the early key events associated with 1,4-DX is limited. There is, however, only limited identified direct  evidence of the essentiality of the key events in the hypothesized MOA addressed through knockout models (e.g., for cyp2E1 induction) or reversibility studies (e.g., for hepatotoxic effects), as noted also by Health Canada (2018).  There is also limited evidence of consistency of the hypothesized mode of action in the core reference material identified for this review based on, for example, knowledge of species and strain variations in cyp2E1 induction.  There is, however, some evidence of analogy with the results of QSAR analysis predicting that 1,4-DX would not induce mutagenicity in Salmonella, chromosomal aberrations in Chinese hamster ovary cells, or unscheduled DNA synthesis in rat hepatocytes but would induce sister chromatid exchange in Chinese hamster ovary cells, micronuclei in the bone marrow of rats and cancer in rats and mice (Rosenkranz and Klopman, 1992 in Health Canada, 2018).
  
  For empirical support, the review materials identified for this poll focused principally on dose-response concordance across hypothesized late key events (often with separate consideration of temporal aspects), a function in part of the limited available data on the nature of and dose-response for early key events. None included integrated analyses of dose-response and temporal concordance for key events for hypothesized MOAs based on benchmark doses (BMDs), though many of the relevant BMD/Cs for late key events were developed and considered separately for cancer and non-cancer effects by US EPA (2013).  The most extensive and recent analyses are those of Dourson et al. (2017) and Lancanfroni et al. (2023). Though challenging due to varied study designs, Dourson et al. (2017) provided helpful graphical illustration of the comparative incidence of the hepatic precursor and neoplastic lesions in the key carcinogenesis bioassays and related subchronic studies for the late key events in the hypothesized MOA (namely, hepatic toxicity, regenerative proliferation and tumours) based on translated versions of the full study reports for the key Japanese carcinogenesis bioassays in rats and mice and a reread of the slides for hepatic effects in rats and mice in NCI studies (JRBC, 1990a, 1990b, NCI, 1978, McConnell reread of NCI, 1978). Lacanfroni et al. (2023) extended analyses to include results from more recent mechanistic studies including transcriptomic analyses relevant to earlier key events in the hypothesized mode of action (Lacanfroni et al., 2021; Chappell et al., 2021; Charkoftaki et al., 2021; Chen et al., 2022).
  
  Based on the conducted analyses, with few exceptions, the observed patterns of dose-response and temporal concordance broadly support the late key events in the hypothesized mode of action above (e.g., Dourson et al., 2017, Health Canada, 2018; Lafranconi et al., 2023), with associated dose ranges in rats and mice, as follows (data from Dourson et al., 2017 with variations of Lancanfroni et al.,  2023, noted; inconsistent information italicized ): 
   

  rats: metabolic saturation at 30 to 100 mg/kg, cellular swelling, hypertrophy and liver weight increases at 42 to 55 mg/kg-day, necrosis and/or inflammation at 94 to 219 mg/kg-day, increased DNA synthesis at 330 mg/kg-day, hyperplasia development at 55 to 330 mg/kg-day, basophilic and mixed cell (when measured) foci development at 55 to 389 mg/kg-day and adenomas and carcinomas at doses of 66 to 1015 mg/kg-day (> 30 mg/kg bw-day in Lancanfroni et al., 2023). 
    

  Mice: metabolic saturation at ~200 mg/kg – day, cellular swelling, hypertrophy and liver weight increase, in the range of 190 to 200 mg/kg-day, necrosis and/or inflammation in the same range, hyperplasia at doses as low as 380 mg/kg-day in one study but not the other  and foci development at doses as low as 380 mg/kg-day in one study but not the other  and adenomas and carcinomas at doses of 66 to 1015 mg/kg-day (200 – 400 mg/kg-day in Lancanfroni et al., 2023) 

  
  One of the important outliers noted in the italicized text above is the lack of observation of later hepatic effects (i.e., proliferation) prior to tumour development in mice in the Japanese carcinogenesis bioassay, which was influential in US EPA’s (2013) considering the data inadequate to support the hypothesized cytotoxic mode of action of 1,4-DX in liver.  Dourson et al. (2017) and Lacanfroni et al. (2023) note, as well, the seeming inconsistency with the observation of non-neoplastic liver pathology in the 13-week drinking water study in mice conducted by the same investigators (Kano et al., 2008) as well as the increases in liver enzymes in the cancer bioassay, expected to be associated with adverse hepatic effects. 
   
  Recent studies incorporating genomic analyses prompted in part by these seemingly discrepant observations have provided additional data on likely early key events in the mode of induction of tumours by 1,4-DX (Lacanfroni et al., 2021, Chappell et al., 2021; Charkoftaki et al., 2021; Chen et al., 2022).  This evolving evidence is consistent with the pattern of results for 1,4-DX in genotoxicity studies in vitro and in vivo.  However, data relevant to consideration of empirical support for hypothesized early key events such as oxidative stress and DNA damage in this dataset are limited, due to the paucity of dose-response information available from the studies (with effects having been observed generally only at the highest doses associated also with cyp 2E1 induction).
   
  For example, Lafranconi et al. (2021) evaluated both the dose-response and time course of hepatic events in female B6D2F1 mice treated with 20, 40, 200, 600, 2000 or 6000 ppm 1,4-DX in drinking water for 7, 28 or 90 days. Liver weight increases after 90 days of exposure were accompanied by evidence of increased pan-lobular hepatocellular proliferation as determined by increased BrdU incorporation at the top dose. Other than limited evidence of single-cell necrosis typical of apoptosis, there was no histological or biochemical evidence of cytotoxicity at any of the exposures used in this study. There was evidence of changes in genomic signaling only at 2000 ppm (337–391 mg/kg/d) and 6000 ppm (895–1063 mg/kg/d) from whole transcriptome analyses consistent with mitotic events (Chappell et al., 2021).  
   
  Based on histopathological and transcriptomic analyses of mouse liver, Charkoftaki et al. (2021) reported alterations in oxidative stress response accompanying DNA damage following 4 weeks exposure to high doses of 1,4-DX (5,000 ppm) in drinking water based on the enrichment of Nrf2and GSH mediated pathways in transcriptomic analyses. In follow up investigation, this group also provided some evidence of the essentiality of oxidative stress as a key event.  Following exposure for one week  (1000 ppm) or 90 days (5000 ppm) in wild type and GSH-deficient glutamate-cysteine ligase modifier subunit (Gclm)-null mice, at 90 days, mild liver cytoxocity, progressive CYP2E1 induction, oxidative stress  (persistent NRF2 induction), oxidation of the GSH pool and accumulation of the lipid peroxidation by-product 4-hydroxynonenal, and elevations in oxidative DNA damage and DNA repair response were observed. These DX-elicited changes were exaggerated in GSH-deficient mice (Chen et al., 2022). 
   
  Collectively, then, available data support an indirectly genotoxic mode of action with early key events, as indicated above - i.e., induction of cyp 2E1,  oxidative stress, DNA damage, hepatoxicity, sustained proliferation and tumours.  Uncertainty is greatest for the early key events between the induction of cyp 2E1 and hepatotoxicity, for which there is limited evidence for 1,2-DX of empirical support (temporal and dose-response concordance) but some experimental support for the essentiality of oxidative stress. 
• View answer

  Expert 2

  	-5	-4	-3	-2	-1	0	1	2	3	4	5
  Direct genotoxicity	0	0	0	0	0	1	0	0	0	0	0
  Indirect genotoxicity	0	0	0	0	0	0	0	0	0	0	1
  Proliferative regeneration induced by cytotoxicity	0	0	0	0	0	0	0	0	0	0	1
  Oxidative stress	0	0	0	0	0	0	0	0	0	0	1
  Activation of nuclear receptors and associated transcription factors (e.g., CAR/PXR)	0	0	0	0	0	0	0	0	1	0	0
  Other (please explain)	0	0	0	0	0	0	0	0	0	0	1

  The onset of hepatic DNA damage is evidenced by DSBs with gH2AX accumulation and unscheduled DNA synthesis as elicited by the BrdU or 3H-thymidine tracers. These culminate in hepatic growth-arrest as demonstrated by the accumulation of highly polyploid hepatocytes with meganuclei.  This is noted in Chen et al. 2022 (Fig. 1B showing H&E images - note panels on the extreme right).  The polyploid cells contain extensive DSBs with gH2AX expression as noted in Chen et al. 2022 (Fig. 5C shows H2AX-positive polyploid cells in panels on the extreme right). 
  
  The designation of "proliferative regeneration" is incorrect because the studies of DNA synthesis in situ through the above markers are insufficient for establishing cell division.  The onset of mitochondrial DNA damage leading to their inability for fission followed by the transfer to daughter cells is another mechanism for interruptions in cytokinesis leading to polyploidy, but this aspect has not been studied. 
  
  The likelihood is that when hepatocytes are damaged by excessive oxidative stress, this will restrict these from cycling, leading to the origin and proliferation of alternative cell populations substantially at risk for genetic mutations contributing to oncogenesis and liver tumors. Such a mechanism will be similar to other major causes of liver cancers, including alcohol-associated steatosis, metabolic dysfunction-associated steatohepatitis (MASH), chronic hepatitis B and C, hemochromatosis, hereditary tyrosinemia, etc. 
  
  The role of cell interactions in this process has not been developed. The bulk transcriptomics and other omics assays of whole liver lysates are insufficiently informative because of the dilution of key events in DNA damage response, cell cycling, and cell survival events by various cell types.        
• View answer

  Expert 3

  	-5	-4	-3	-2	-1	0	1	2	3	4	5
  Direct genotoxicity	1	0	0	0	0	0	0	0	0	0	0
  Indirect genotoxicity	0	0	1	0	0	0	0	0	0	0	0
  Proliferative regeneration induced by cytotoxicity	0	0	0	0	0	1	0	0	0	0	0
  Oxidative stress	0	0	0	0	1	0	0	0	0	0	0
  Activation of nuclear receptors and associated transcription factors (e.g., CAR/PXR)	0	1	0	0	0	0	0	0	0	0	0
  Other (please explain)	0	0	0	0	0	0	1	0	0	0	0

  There is little/no evidence of direct genotoxicity certainly in vitro and only high dose related activity (above tumorigenic doses); however, there may be at best a very contribution of oxidative stress.  However, caution must be exercised in attributing oxidative stress to a genotoxic carcinogenic MoA in that relatively potent oxidative stress inducers (e.g., paraquat and diquat, oxidative stress caused by redox cycling) are not rodent carcinogens.  However, recent transcriptomic data provides supports for a potentially strong mitogenic liver cell proliferation, and CAR/PXR responses are not consistent with responses associated with known NR liver carcinogens, e.g., phenobarbital.  Metabolism/saturation of cyp2E1 appears very critical to outcome in that DX, and suggests a contributing (but not necessarily a genotoxic) role of oxidative stress associated with cyp2e1 metabolism/induction (supported by attentuation of liver toxicity and oxidative stress in cyp2e1 knockout mice).