<p> This dissertation includes three separate but related studies performed in partial fulfillment of the requirements for the degree of Doctor of Public Health in Environmental and Occupational Health. The main goal this dissertation was to develop and assess quantitative relationships for predicting doses associated with chronic non-cancer toxicity levels in situations where there is an absence of chronic toxicity data, and to consider the applications of these findings to chemical substitution decisions. Data from National Toxicology Program (NTP) Technical Reports (TRs) (and where applicable Toxicity Reports), which detail the results of both short-term and chronic rodent toxicity tests, have been extracted and modeled using the Environmental Protection Agency&rsquo;s (EPA&rsquo;s) Benchmark Dose Software (BMDS). Best-fit minimum benchmark doses (BMDs) and benchmark dose lower limits (BMDL) were determined. Endpoints of interest included non-neoplastic lesions, final mean body weights and mean organ weights. All endpoints were identified by NTP Pathologists in the abstract of the TRs as either statistically or biologically significant. A total of 41 chemicals tested between 2000 and 2012 were included with over 1700 endpoints for short-term (13 week) and chronic (2 year) exposures. </p><p> Non-cancer endpoints were the focus of this research. Chronic rodent bioassays have been used by many methodologies in predicting the carcinogenic potential of chemicals in humans (1). However, there appears to be less emphasis on non-cancer endpoints. Further, it has been shown in the literature that there is little concordance in cancerous endpoints between humans and rodents (2). The first study, Quantitative Relationship of Non-Cancer Benchmark Doses in Short-Term and Chronic Rodent Bioassays (Chapter 2), investigated quantitative relationships between non-cancer chronic and short-term toxicity levels using best-fit modeling results and orthogonal regression techniques. The findings indicate that short-term toxicity studies reasonably provide a quantitative estimate of minimum (and median) chronic non-cancer BMDs and BMDLs. </p><p> The next study, <i>Assessing Implicit Assumptions</i> in Toxicity Testing Guidelines (Chapter 3) assessed the most sensitive species and species-sex combinations associated with the best-fit minimum BMDL10 for the 41 chemicals. The findings indicate that species and species-sex sensitivity for this group of chemicals is not uniform and that rats are significantly more sensitive than mice for non-cancerous outcomes. There are also indications that male rats may be more than the other species sex groups in certain instances. </p><p> The third and final study, <i>Comparing Human Health</i> Toxicity of Alternative Chemicals (Chapter 4), considered two pairs of target and alternative chemicals. A target is the chemical of concern and the <i>alternative </i> is the suggested substitution. The alternative chemical lacked chronic toxicity data, whereas the target had well studied non-cancer health effects. Using the quantitative relationships established in Chapter 2, Quantitative Relationship of <i>Non-Cancer Benchmark Doses in Short-Term and Chronic Rodent Bioassays,</i> chronic health effect levels were predicted for the alternative chemicals and compared to known points of departure (PODs) for the targets. The findings indicate some alternatives can lead to chemical exposures potentially more toxic than the target chemical.</p>

<p> Non-alcoholic fatty liver disease (NAFLD) has become the leading cause of chronic liver disease. The spectrum of NAFLD ranges from simple steatosis, to non-alcoholic steatohepatitis (NASH), fibrosis, cirrhosis, and potentially hepatocellular carcinoma. Dietary factors and chemical exposure have been associated with the disease progression. In addition, the presence of NAFLD changes the metabolism of drugs and chemicals, which may in turn increase the susceptibility of the liver to xenobiotic induced toxicity. To examine the potential interplay of chemicals on diet-induced NAFLD, three studies were conducted in this dissertation project. In the first study, a mouse model was established that recapitulated the spectrum of liver damages seen in human NAFLD. Using a high fat diet (HFD), steatosis, NASH, progressive fibrosis, and liver tumor formation were produced in mice. Modulations of nuclear receptors involved in metabolism of endogenous and xenobiotic compounds were characterized at various stages of NAFLD. Using this mouse model, a second study examined if perfluorooctanoate (PFOA), a ubiquitous environmental contaminant, modulated the progression of NASH. The results showed PFOA induced hepatic DNA synthesis and liver inflammation were exacerbated in the mice fed with HFD. In contrast, PFOA decreased the severity of hepatic steatosis and fibrosis induced by HFD. To further investigate the mechanisms underlying these observed effects, a third study was performed that analyzed the hepatic transcriptome in liver samples taken from studies 1 and 2. The results of the third study demonstrated that cytokine and chemokine-related genes played important role in the development of both inflammation and fibrosis in NAFLD. Comparing PFOA to vehicle controls in HFD fed mice, PFOA disrupted the lipid homeostasis to favor clearance in the fatty liver, as most of the genes were enriched in the fatty acid oxidation pathways. In summary, this project established and a mouse model of HFD induced NAFLD and characterized the interplay of diet and chemicals in the disease progression. The results from this dissertation also indicated that patients with NAFLD may respond differently compared with healthy individuals. The potential susceptibility of this population to chemically induced hepatotoxicity needs to be carefully considered when assessing risk.</p><p>

<p> Millions of people are exposed to arsenic in the United States and worldwide. Commonly found arsenic species in human urine are AsIII (arsenite), AsV (arsenate), MMA (monomethyl arsenic acid), DMA (dimethylarsinic acid) and AB (arsenobetaine). Evidence has shown that these species vary in toxicity, and since each of these metabolites can be detected through analysis, they have the potential to be used as biomarkers for human exposure. For human exposure assessments in areas that have naturally occurring arsenic contaminated sources, or those who live or work near contaminated environmental sites where arsenic has been used, it is important to fully understand what species of arsenic residents are being exposed to in order to grasp the risk of arsenic exposure specifically and in its entirety.</p><p> Since it is difficult to determine direct human exposures, a swine model was used as a surrogate. Swine urine was collected from two different swine studies where animals were given non-toxic doses of arsenic contaminated soil and another group receiving a soluble reference dose using sodium arsenate for comparison. The urine samples from these studies were used to modify an arsenic speciation method using high-performance liquid chromatography and inductively coupled plasma mass spectrometry (LCICPMS). It is evident that when comparing the percent of arsenic species found in swine urine samples with what is found in humans a correlation can be made. There was a range of 64&ndash;74% DMA in swine samples for all test soils where a range of 60&ndash;75% DMA has been reported in human urine samples. This further illustrates the importance of arsenic speciation in swine urine since it does appear that it could correlate to human exposure. If proper measurement systems are utilized to quantify As species of health concern, dosed swine can be used to assess and predict human toxicological effects of arsenic exposure.</p>

<p>Endocrine disrupting compounds (EDCs) from agricultural, industrial, and municipal sources can be found in many surface waters with potential adverse implications for human and ecosystem health. The South Platte River represents a significant source of water for the Denver Metro Area, yet little data exists concerning EDCs. The aim of this study is to evaluate the occurrence and effects of EDCs downstream from two major wastewater treatment plants (WWTPs). This study characterizes the estrogenicity gradient of the South Platte River in the Denver Metro area by combining data from qPCR analysis for liver vitellogenin (<i>vtg</i>) mRNA with liver NMR metabolomics after a 5 day in situ caged exposure of fathead minnows. Concurrent water samples collected from the start and end times of the exposures were used to determine the occurrence and concentration of wastewater contaminants. Results found 68 of 122 chemicals downstream of WWTP 1 and 73 downstream of WWTP 2, including known EDCs (e.g. nonylphenol and octylphenol). A steroidal estrogen, estrone, was only found downstream of WWTP 2. Consistent with the highest measured concentrations of wastewater estrogens, the highest levels of <i>vtg</i> mRNA were measured downstream of WWTP 2. Metabolomics data coincided with <i>vtg </i> data and showed little variation except downstream of WWTP 2, where male polar metabolomes showed increased levels of alanine and glutamate, which are utilized in VTG synthesis. PCA of male polar metabolomes showed significant separation of WWTP 2 from WWTP 1 and the reference site, further supported by PLS-DA scores plot. Female polar metabolomes showed significant separation between WWTP 1 and WWTP 2 using PLS-DA scores plot. This study demonstrates that qPCR and metabolomics data can be reliably and concurrently used to illuminate impacts from chemical exposures, although further research will better elucidate target genes and metabolites of interest. </p>