Background
Triphenyl phosphate (TPHP) is an organophosphate flame retardant currently on the market and used as a replacement for phased-out polybrominated diphenyl ethers. Toxicity information on this class of chemicals is inadequate. Herein, we use short-term, in vivo transcriptomic studies to inform potential points of departure for TPHP.
Methods
The National Toxicology Program (NTP) conducted a transcriptomic study on TPHP, in male Harlan Sprague Dawley rats. TPHP was dissolved in corn oil. Exposure was once daily for 4 days by oral gavage. TPHP (>99%) was tested at six doses (0, 55, 110, 220, 441, and 881 mg/kg body weight). On Day 5, animals were sacrificed, standard toxicological measures assessed, and the liver taken for gene expression studies using Affymetrix microarrays. Modeling was conducted to identify the benchmark doses (BMDs) associated with the most sensitive apical toxicological endpoints and with transcriptional changes in the liver at a benchmark response of one standard deviation from the mean.
Results
The most sensitive apical endpoints for which BMD values could be obtained were serum HDL (high-density lipoprotein) cholesterol levels, absolute liver weights, relative liver weights, and serum cholesterol levels. The benchmark dose lower confidence limit BMDLs (and BMDs) were 39 (79), 48 (136), 71 (103), and 90 (142) mg/kg, respectively. Although serum cholinesterase appeared to be a sensitive endpoint (35–70% decrease) at all doses, beginning with 55 mg/kg (the lowest-observed-effect level), its BMD could not be determined due to poor model fit.
Sensitive transcriptional gene set changes by potency included 14 Gene Ontology Biological Processes with BMD median values below the lower limit of extrapolation (<18.3 mg/kg). The most sensitive gene sets for which a reliable estimate of the BMD could be made were cellular polysaccharide biosynthetic process and oligodendrocyte development with median BMD of 19 mg/kg and BMDL medians of 11 mg/kg for both gene sets. The most potently affected single genes ranking in the top 10 for increased fold expression change respective to control were Ces2c and Cyp2b1 with BMDs <18.3 mg/kg (the lower limit of extrapolation, 3× lower than the lowest tested dose) and maximal fold increases of 16.3 and 10.5, respectively. The most potently affected single genes ranking in the top 10 for decreased fold expression change respective to control were Scd and G6pc with BMD of <18.3 mg/kg and BMDL of 16 mg/kg, respectively, and maximal fold decreases of −11.1 and −5.1, respectively.
Conclusion
Taken together, the most sensitive BMD gene set medians that could be reported occurred at 19 mg/kg (BMDL 11 mg/kg) and the most sensitive apical endpoint was increased serum HDL cholesterol, with a BMDL of 39 mg/kg. Cholinesterase inhibition was observed, with effects occurring at all doses including the lowest tested dose of 55 mg/kg. Future studies investigating lower doses would be helpful to obtain more accurate estimates of BMD values for the most sensitively affected genes and for cholinesterase inhibition.
National Toxicology Program. 2018. NTP research report on in vivo repeat dose biological potency study of triphenyl phosphate (CASRN 115-86-6) in male Sprague Dawley rats (Hsd: Sprague Dawley SD) (gavage studies). NTP RR 8. Research Triangle Park, NC: National Toxicology Program (8): 1-36. https://doi.org/10.22427/NTP-RR-8