What additional follow-up (e.g., additional studies, PBPK model refinement) would increase confidence in the hypothesized mode of action for rat nasal cavity squamous cell carcinoma and rat peritoneal mesothelioma (i.e., what, if any, constitute highest priorities for follow-up work)?

Additional mechanistic evidence to identify key events, their essentiality and empirical support (dose-response and temporal concordance).  (See response to 2.1).  Collection of additional toxicokinetic data in mechanistic studies and refinement of the PBK models, if possible, to address US EPA concerns.

The incorporation of risks for somatic mutations in modeling will be helpful. The conclusions regarding proliferative regeneration based on DNA synthesis assays using nucleotide tracers are problematic - this will benefit from cell transplantation assays.    

Such efforts are unlikely to improve confidence defining the human dosimetric relevance of rodent nasal tumors in that defining the systemically-contributed vs site-of-contact dosimetry as key mechanistic drivers of the response seems intuitively very challenging (and thus a low-priority effort for enhancing clarity of any hypothesized MoA).

For nasal tumors, there is some evidence that oxidative stress can be causal. An interesting example is wood dust (a human nasal carcinogen) which may cause oxidative stress upon sustained invasion of inflammatory cells (Bono et al., Environ Res.2019). Although a number of DNA-reactive carcinogens can cause nasal tumors, it is also reported that oxidative stress (nickel; Guo et al., Int J Mol Sci. 2019  ) typically in combination with cytotoxicity and regenerative proliferation (formaldehyde; Thompson et al., Crit. Rev. Toxicol 2020) are plausible mechanisms leading to nasal tumors. Thus, I conclude that the same MoA established for liver tumors may also be responsible for the nasal tumors seen with DX. 

The mesothelioma of the rat peritoneum (typically the processus vaginalis testis in males) is seen spontaneously, e.g., in Fischer rats but also upon treatment with chemical carcinogens. It is a rare tumor in humans. Nevertheless it is occasionally used as a critical endpoint since the chemicals often have a genotoxic potential and often cause tumors in other organs as well. 

Typical examples a bromochloro acetic acid and o-nitrotoluene (Kim et al., TAP 2006). The latter can lead to oxidative DNA damage and  covalently bound DNA adducts (Watanabe et al., Environ Health Prev Med. 2010; Jones et al., Carcinogenesis 2003 ), while the former causes oxidative DNA damage (Austin et al., Fund. Appl. Toxicol. 1996). Thus, oxidative stress is likely to be a major trigger of DNA damage and peritoneal mesothelioma formation in rats. It appears plausible that such a mechanism can also lead to tumors caused by DX.

I am not sure if PBPK modeling will help define the MOA for these two tumor sites.  Suggested additional studies would include a short term in vivo study in 1,4 DX treated rats are several doses (carcinogenic doses and subcarcinogenic dose).  This study should include several sampling times (1 wk, 2wk 4 wks ) after continuus exposure. Endpoints to consider would be histopathlogy (including measurement of cell proliferation  and apoptosis), gene expresion (including nuclear receptor gene expression (ie CAR, PXR, AHr, Estrogen etc), oxidative stress/damage, etc .  These two later endpoints may be difficult to examine given the amount of tissue available and thus the histopathology may be the most important and most easily obtainable endpoint to address initially.  Based on the results of this short term in vivo study, additional in vitro studies could be performed to dissect the mechanism involved.