Tunicamycin challenge elicits ER stress and CHOP induction in male and female mice without significant cell death.

(A) Whole livers from male or female mice challenged for the indicated times with 1 mg/kg b.w. TM or vehicle (veh) (B) Immunoblot to detect CHOP and the ER-resident glycoprotein TRAPα, which exists normally as a doubly glycosylated form (TRAPα**) and for which mono-glycosylated (TRAPα*) and unglycosylated (TRAPα) forms are induced by TM. Calnexin (CNX) was used as a loading control. Asterisk under lane 17 designates an animal that did not receive a fully effective dose of TM, as indicated by very modest inhibition of TRAPα glycosylation. Here and elsewhere, hairlines are used only for visual clarity. (C) TUNEL staining was used to detect cell death at the indicated time points, with 300 mg/kg acetaminophen for 24h used as a positive control.

Hepatocyte-specific deletion of CHOP attenuates IRE1 activation and expression of ISR-regulated genes, but not XBP1- or ATF6-regulated genes.

(A) Schematic showing the Chop FLuL allele, in which CHOP expression is silenced by a gene trap; the floxed allele, in which CHOP expression is wild-type by virtue of FLPo-mediated excision of the gene trap; and the deleted allele, in which the entire open reading frame of CHOP is deleted by CRE. Also shown is the nTnG CRE reporter allele bred into Chopfl/fl mice, which converts from expression of tdTomato to EGFP upon CRE action. Figure adapted from (Liu et al. 2024). (B) Experimental paradigm for Chop deletion in Chopfl/fl; nTnG animals and challenge with 1 mg/kg TM (C) tdTomato, EGFP, and nuclei (DAPI) were visualized, and CHOP detected by immunostaining, in livers of an animal with an intact Chop allele expressing nuclear tdTomato in all cells and cytoplasmic AAV-driven EGFP in hepatocytes (top row), and an animal with a deleted Chop allele expressing nuclear tdTomato in all non-hepatocytes and nuclear EGFP in hepatocytes (bottom row). Note the absence of CHOP expression in the liver when CHOP is deleted specifically in hepatocytes. (D) In situ liver images from non-deleted (fl/fl) or hepatocyte-deleted (HKO) animals after challenge (E) Direct measurement of liver triglyceride normalized against protein content. (F) Immunoblot showing inhibition of glycosylation by TM using TRAPα as an indicator, as well as expression of the ER chaperone and UPR target gene BiP and phosphorylated and total species of eIF2α. Consistent with Figure 1, CHOP itself is undetectable in unchallenged livers. (G) Quantification of eIF2α phosphorylation from (F) (H) qRT-PCR detection of the indicated ISR target genes (I) Xbp1 mRNA splicing detected by conventional RT-PCR with primers that equivalently recognize both spliced and unspliced forms (J, K) qRT-PCR detection of targets of the IRE1-XBP1 axis (J), a canonical RIDD target (Bloc1s1; J), and targets of the ATF6 axis (K). Here and elsewhere, error bars represent means +/- SDM and statistical comparisons were by two-way ANOVA.

RNA-seq reveals a broad effect of CHOP in augmenting stress-dependent gene regulation.

(A) Principal component analysis of samples used for RNA seq. ChopFLuL/FLuL animals were exposed to AAV-TBG-GFP or AAV-TBG-FLPo; the former treatment leaves the animals null for CHOP in all tissues (ChopKO), while the latter restores CHOP competence specifically in hepatocytes (w.t.Hep). Genotype is indicated by color, sex by shape, and TM treatment by fill. (B) Xbp1 RNA splicing from a larger cohort of animals including the 16 used for RNA-seq (C) A heatmap showing all genes significantly regulated (FDR<0.05) more than 2-fold by tunicamycin in w.t.Heplivers, ordered by magnitude of regulation with upregulated genes on top (D, E) Pathway analysis of genes that were lower (D) or higher (E) in their expression in TM-treated ChopHKO livers compared to wild-type TM-treated. (F) Genes from (C) were plotted for log10 induction by TM in w.t.Hep versus ChopKO livers. Dashed black line shows the best-fit line and associated equation while the dashed red line shows a hypothetical fit if CHOP deletion had no effect on gene expression. Colored quadrants show genes different in their expression 1.5-fold or greater compared to expected expression from the equation. Pie charts show percentage of genes directly bound by CHOP (from ChIP-seq analysis in Figure 4) or not bound in groups (a) and (d) with all 7 directly bound genes in group (a) listed below. (G) Pathway analysis of genes from groups (a), (b), and (c+d) in panel F

ChIP-seq identifies liver-specific CHOP target genes.

(A) Genome-wide frequency of CHOP binding to the indicated DNA regions, with promoters defined as within 3 kb of a transcriptional start site (B) The most enriched known motif binding sites for CHOP genome-wide (C) Overlap between promoters bound by CHOP and genes significantly more highly expressed in w.t.Hep livers after TM challenge than in ChopKOlivers (D) Overlap between the 208 bona fide CHOP targets for upregulation shown in (C) and similarly identified targets in Han et al.(Han et al. 2013) (E) Pathway analysis of bona fide targets from (C) (F) Motif analysis of bona fide targets from (C) (G-J) Similar to (C-F) except for CHOP targets for suppression (K) ChEA3 analysis to identify enriched transcription factor binding sites among genes in group (a) from Figure 3F

Weakening of CHOP-dependent regulation in primary hepatocytes in vitro.

(A-C) Select newly-identified bona fide targets of CHOP were analyzed by qRT-PCR for their expression in the liver after an 8h TM challenge (A), or in Chopfl/fl primary hepatocytes treated with Ad-CRE to delete Chop or Ad-GFP as a control, after an 8h challenge with 5 μg/ml TM (B) or 500 nM TG (C). Gray boxes indicate genes that do not show similar significant CHOP-dependent differences after ER stress challenge in vitro as in vivo. (D-F) Similar to (A-C), except examining genes involved in hepatocyte identity that are not direct targets of CHOP (G, H) Xbp1 mRNA splicing after in vitro TM or TG treatment as above

Hepatocyte deletion of ATF6α exacerbates ER stress and permits eIF2α-independent ISR signaling.

(A) Experimental paradigm for Atf6α deletion in Atf6αfl/fl animals and challenge with 1 mg/kg TM for 48h. (B) Xbp1 mRNA splicing (C) Whole livers (D) Hematoxylin and eosin staining showing extensive microvesicular steatosis in Atf6αHKO livers after challenge (E) Immunoblot to detect CHOP, phosphorylated and total eIF2α, and TRAPα(F-I) qRT-PCR quantification of the indicated genes including Atf6 itself (F), ISR targets (G), IRE1-XBP1 and RIDD targets (H), and hepatocyte identity genes (I) (J) Detection of ATF4 in purified nuclei, along with a total protein stain

Simplified schematic figure depicting the role of CHOP in mediating the transition from first-phase to second-phase stress signaling.

See discussion for details.