Identifying TCDD-resistance genes via murine and rat comparative genomics and transcriptomics

The aryl hydrocarbon receptor (AHR) mediates many of the toxic effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). However, the AHR alone is insufficient to explain the widely different outcomes among organisms. Attempts to identify unknown factor(s) have been confounded by genetic variability of model organisms. Here, we evaluated three transgenic mouse lines, each expressing a different rat AHR isoform (rWT, DEL, and INS), as well as C57BL/6 and DBA/2 mice. We supplement these with whole-genome sequencing and transcriptomic analyses of the corresponding rat models: Long-Evans (L-E) and Han/Wistar (H/W) rats. These integrated multi-species genomic and transcriptomic data were used to identify genes associated with TCDD-response phenotypes. We identified several genes that show consistent transcriptional changes in both transgenic mice and rats. Hepatic Pxdc1 was significantly repressed by TCDD in C57BL/6, rWT mice, and in L-E rat. Three genes demonstrated different AHRE-1 (full) motif occurrences within their promoter regions: Cxxc5 had fewer occurrences in H/W, as compared with L-E; Sugp1 and Hgfac (in either L-E or H/W respectively). These genes also showed different patterns of mRNA abundance across strains. The AHR isoform explains much of the transcriptional variability: up to 50% of genes with altered mRNA abundance following TCDD exposure are associated with a single AHR isoform (30% and 10% unique to DEL and rWT respectively following 500 μg/kg TCDD). Genomic and transcriptomic evidence allowed identification of genes potentially involved in phenotypic outcomes: Pxdc1 had differential mRNA abundance by phenotype; Cxxc5 had altered AHR binding sites and differential mRNA abundance. Author Summary Environmental contaminants such as dioxins cause many toxic responses, anything from chloracne (common in humans) to death. These toxic responses are mostly regulated by the Ahr, a ligand-activated transcription factor with roles in drug metabolism and immune responses, however other contributing factors remain unclear. Studies are complicated by the underlying genetic heterogeneity of model organisms. Our team evaluated a number of mouse and rat models, including two strains of mouse, two strains of rat and three transgenic mouse lines which differ only at the Ahr locus, that present widely different sensitivities to the most potent dioxin: 2,3,7,8 tetrachlorodibenzo-p-dioxin (TCDD). We identified a number of changes to gene expression that were associated with different toxic responses. We then contrasted these findings with results from whole-genome sequencing of the H/W and L-E rats and found some key genes, such as Cxxc5 and Mafb, which might contribute to TCDD toxicity. These transcriptomic and genomic datasets will provide a valuable resource for future studies into the mechanisms of dioxin toxicities.

236 liver or TCDD-sensitive cohorts (Fig 2D-E). Genes with altered mRNA abundance in 237 DBA/2 mice included Apol7c, Tnfaip8l3 and Htatip2 (both low and high dose 238 exposure). These genes were similarly altered in the sensitive strains. Rpl18a and 239 Mbd6 had altered mRNA abundance exclusively in the resistant group (high and low 240 dose), while two additional genes, Onecut2 and Lipg had altered mRNA abundance 241 exclusively in the resistant mouse, high dose group (Fig 2D) 269 utilizing a dose-response experiment. As above, we identified a large number of 270 genes with differential mRNA abundance in the TCDD-sensitive rWT mouse, as well 271 as the more resistant DEL mouse, and a muted response in highly resistant INS 272 mouse (Fig 3A). As a clear dose-response regarding number of differentially 273 abundant RNAs was not apparent, we focused downstream analyses to the 500 274 µg/kg TCDD group, for consistency with EXP1. As explained above, this dose is   Table). Additionally, no genomic differences were detected 355 between the TCDD-sensitive L-E and TCDD-resistant H/W rats (S3 Table), 356 suggesting that Pxdc1 is a poor candidate for the proposed gene "B".
357 Pcp4l1 demonstrated induced RNA abundance in most cohorts (Fig 4C, top). 367 response following TCDD exposure across species (Fig 4C, bottom): hepatic mRNA 368 abundance was increased after TCDD exposure in C57BL/6 mice (4 days, all doses) 369 and earlier in rWT mice and L-E rats (19 hours). This response was followed by a 370 significantly reduced mRNA abundance in rWT mice (4 days) and L-E rats (4 and 10 371 days). This gene also shows different presence of AHREs in its promoter region 372 between species (n = 2 AHRE-1 (core) motif in both mice and rats; 4 ARE motifs in 373 mice; 1 AHRE-2 and 1 ARE motif in rats). No differences in AHREs were observed 374 between H/W and L-E rats for either Pcp4l1 or Nr1i3. This makes these genes 375 interesting candidates for involvement in TCDD-induced toxicity.
377 datasets, using only orthologous genes (S4 Table). Unsurprisingly, sensitive strains 378 exhibit a larger number of significantly enriched pathways than do resistant animals, 379 many of which are altered in multiple strains/lines (Fig 4D).  Table), 209 mapped to genes evaluated in our microarray cohorts. 416 We next evaluated differential transcriptional patterns for genes demonstrating 417 differences among these TFBSs between H/W and L-E rats. For example, Cxxc5 418 shows a loss of the AHRE1 (full) motif within its promoter region in H/W (one found in 419 H/W and two in L-E/rn6) and exhibits significantly reduced transcript abundance in 420 liver from both L-E (4 and 10 days) and the TCDD-sensitive Ln-C rat (19 hours) 421 following exposure to TCDD (Table 2)    490 genes may be due to additional transcription factors as either primary or secondary 527 mRNA abundance for this gene was significantly increased in TCDD-exposed liver 528 from L-E rats (4 and 10 days, log 2 fold change = 1.3 and 1.7, p adj = 1.16x10 -12 and 529 1.16x10 -17 ) with a more muted response in H/W rats (4 and 10 days, log 2 fold change 530 = 0.57 and 0.81, p adj = 3.7x10 -4 and 4.28x10 -8 ). These genes provide interesting 531 candidates for gene "B" that require further studies into its potential involvement in 532 the onset of TCDD-toxicities.
533 The purpose of this study was to ascertain the mechanism of classic TCDD toxicity 534 using various model systems, including transgenic mice to compare various rat Ahr 535 variants in a system with a homogeneous genetic background, and various strains of 536 rat, each with differing phenotypic responses to TCDD. To accomplish this, we 537 generated unique transcriptomic and genomic datasets that provide multiple levels of 538 evidence. Using this valuable resource, we identified several genes whose 539 transcription was selectively altered by TCDD in either TCDD-sensitive or TCDD-540 resistant cohorts, a differential response that can be attributed to the particular AHR 541 isoform expressed in each cohort. Pxdc1 in particular demonstrated differential 542 transcription between TCDD-sensitive and TCDD-resistant models across both mice 543 and rats. However, the transcriptional responsiveness of this gene could not be 544 explained by genomic differences in AHR-binding sites, as the transcription factor 545 binding site analysis revealed highly variant sites between species, and no major 546 difference between strains of rat. However, genomic sequence analysis allowed 547 identification of differences between sensitive and resistant rat strains, which are 548 potential "gene B" candidates. For instance, Cxxc5 was found to have fewer 549 occurrences of AHRE-1 (full) TFBSs in H/W relative to L-E, and had reduced RNA 550 abundance in sensitive strains. This is a suitable candidate for further study in 551 relation to mechanisms of TCDD toxicity and regulatory roles of the AHR.

Materials and Methods
553 Animal handling 554 Three separate experiments were performed (Fig 1)