Cell death–associated lipid droplet protein CIDE-A is a noncanonical marker of endoplasmic reticulum stress

Secretory protein misfolding has been linked to ER stress and cell death. We expressed a TGrdw transgene encoding TG-G(2298)R, a misfolded mutant thyroglobulin reported to be linked to thyroid cell death. When the TGrdw transgene was expressed at low level in thyrocytes of TGcog/cog mice that experienced severe ER stress, we observed increased thyrocyte cell death and increased expression of CIDE-A (cell death-inducing DFFA-like effector-A, a protein of lipid droplets) in whole thyroid gland. Here we demonstrate that acute ER stress in cultured PCCL3 thyrocytes increases Cidea mRNA levels, maintained at least in part by increased mRNA stability, while being negatively regulated by activating transcription factor 6 — with similar observations that ER stress increases Cidea mRNA levels in other cell types. CIDE-A protein is sensitive to proteasomal degradation yet is stabilized by ER stress, and elevated expression levels accompany increased cell death. Unlike acute ER stress, PCCL3 cells adapted and surviving chronic ER stress maintained a disproportionately lower relative mRNA level of Cidea compared with that of other, classical ER stress markers, as well as a blunted Cidea mRNA response to a new, unrelated acute ER stress challenge. We suggest that CIDE-A is a novel marker linked to a noncanonical ER stress response program, with implications for cell death and survival.


Introduction
The thyroglobulin (TG) gene is the most highly expressed gene of the thyroid gland, and TG protein makes up more than 50% of all thyroidal protein. In order to serve as substrate for thyroid hormone synthesis, the TG protein must be exported through the secretory pathway (1). Misfolding of mutant TG in the ER is an established cause of congenital hypothyroidism (2), which, via endocrine feedback, triggers an increase in circulating levels of thyroid-stimulating hormone (TSH) released from the anterior pituitary. In this disease (in both humans and animal models), thyroid follicular (thyrocyte) cell growth (3) and thyrocyte cell death (4) take place simultaneously in different cells of the same thyroid gland. This behavior suggests that affected cells in the same tissue exhibit a range of (successful and unsuccessful) attempts in adaptive response to chronic ER stress. Such a finding is consistent with the broader notion that a myriad of prosurvival and prodeath signals are generated as downstream consequences of ER stress (5), and there are likely to be both complex (and possibly subtle) changes in the profile and balance of these activities that trigger death in some cells while allowing survival and even proliferation of adjacent cells of the same cell type.
In the case of the thyroid gland, the net balance of thyroid cell growth and death determines the potential for development of a goiter (which signifies net growth). Thus, in congenital hypothyroidism caused by mutant TG, the ultimate size of the goiter serves as a physiologically integrated readout of the balance of many hundreds of thousands of individual cell fate decisions. One of the striking outcomes of this balance is observed when comparing the congenital goiter (cog/cog) mouse and the dwarf (rdw/rdw) rat, both of which are caused by a single homozygous missense mutation of TG. The mutation in each case resides in the same ChEL (CholinEsterase-Like) domain, just 35 amino acids apart, within the 2765 residue TG protein (Supplemental Figure 1A; supplemental material available online with this article; https://doi. org/10.1172/jci.insight.143980DS1). However, the cog/cog mouse generates an enlarged thyroid (goiter, ref. 6), whereas the rdw/rdw rat develops atrophy of the thyroid gland (7).
Secretory protein misfolding has been linked to ER stress and cell death. We expressed a TG rdw transgene encoding TG-G(2298)R, a misfolded mutant thyroglobulin reported to be linked to thyroid cell death. When the TG rdw transgene was expressed at low level in thyrocytes of TG cog/cog mice that experienced severe ER stress, we observed increased thyrocyte cell death and increased expression of CIDE-A (cell death-inducing DFFA-like effector-A, a protein of lipid droplets) in whole thyroid gland.
Here we demonstrate that acute ER stress in cultured PCCL3 thyrocytes increases Cidea mRNA levels, maintained at least in part by increased mRNA stability, while being negatively regulated by activating transcription factor 6 -with similar observations that ER stress increases Cidea mRNA levels in other cell types. CIDE-A protein is sensitive to proteasomal degradation yet is stabilized by ER stress, and elevated expression levels accompany increased cell death. Unlike acute ER stress, PCCL3 cells adapted and surviving chronic ER stress maintained a disproportionately lower relative mRNA level of Cidea compared with that of other, classical ER stress markers, as well as a blunted Cidea mRNA response to a new, unrelated acute ER stress challenge. We suggest that CIDE-A is a novel marker linked to a noncanonical ER stress response program, with implications for cell death and survival.
JCI Insight 2021;6(7):e143980 https://doi.org/10.1172/jci.insight.143980 In an attempt to understand this difference, our initial animal studies have led us to investigate CIDE-A (cell death-inducing DFFA-like effector-A) (8), which is an established protein of lipid droplets (9,10). Herein, we describe the regulation of Cidea mRNA and protein in PCCL3 (cell line of cultured rat) thyrocytes challenged with acute and chronic ER stress. Specifically, we report that CIDE-A is a gene product linked to a noncanonical acute ER stress response, leading us to consider that CIDE-A may contribute to the complex cell fate program of ER-stressed thyrocytes as well as other cell types.

Results
Identification of Cidea in the thyroid gland of animals experiencing toxic ER stress. Some patients suffering from congenital hypothyroidism with defective TG may develop thyroid gland overgrowth (goiter, ref. 11) whereas others may not. The homozygous TG cog/cog mouse encodes expression of misfolded TG that is defective for export from the ER, triggering thyrocyte ER stress throughout life (12,13) even as the thyroid gland continues to grow. In contrast, the TG rdw/rdw rat expresses homozygous misfolded mutant TG that is also defective for ER export (14,15) accompanied by chronic thyrocyte ER stress (16,17), yet it develops an atrophic thyroid gland (7) with associated thyroid cell death (4). Given that the 2 missense substitutions are so proximate within the TG-ChEL domain, the disparate phenotypes are puzzling. To begin to study the problem, we developed a TG rdw transgene (Supplemental Figure 1, A and B) in which the TG promoter drives thyroid-specific expression (18) of the mutant TG protein (bearing a triple-Myc-epitope tag). Transgenic animals were identified by tail vein PCR (Supplemental Figure 1C); the transgene was expressed at a low level (≤1% of endogenous TG mRNA in TG cog/cog mice) such that total TG mRNA in the thyroid of transgenic animals was unchanged from that in TG cog/cog mice (Supplemental Figure 2A), and both sets of 4-month-old animals had primary hypothyroidism (with circulating T 4 < half of control animals and circulating TSH elevated more than an order of magnitude and no significant differences between them). Nevertheless, thyroidal rdw-TG protein expression was confirmed in transgenic animals (Supplemental Figure 2B). Without the transgene, TG cog/cog animals already exhibited massive ER distension as reported previously (19) -this was also found in the transgenic TG cog/cog mice, as demonstrated by immunofluorescence of the ER marker immunoglobulin heavy chain binding protein (BiP) (Supplemental Figure 2C). Without the transgene, TG cog/cog animals exhibited dramatic ER stress as reported previously (13) -comparable ER stress was also found in transgenic TG cog/cog mice as demonstrated by X-box binding protein-1 (XBP1) splicing ( Figure 1A) with upregulation of ER chaperones, as well as CCAATenhancer-binding protein homologous protein (CHOP) ( Figure 1B). Both sets of TG cog/cog animals exhibited positive TUNEL staining in the thyroid gland, but TUNEL-positive thyroid cell death in the animals bearing the rdw-TG transgene was detectably greater (Figure 2A), and, as measured by live animal thyroid ultrasound ( Figure 2B), expression of the transgene limited the magnitude of thyroid gland growth ( Figure 2C).
In an early analysis, we prepared RNA from TG cog/cog ± TG rdw transgene (n = 4 per group) and performed a simple quantitative PCR (qPCR) microarray of cell death-related mRNAs, accompanied by positive and negative controls. Of the 86 gene products analyzed, the most upregulated mRNA found in the thyroids of TG cog/cog + rdw-TG was Cidea (Supplemental Figure 3A). Although there was considerable variability between individual animals, by qRT-PCR we independently confirmed an average 15-fold elevation of Cidea mRNA in the thyroid glands of transgenic animals ( Figure 2D).
Examination of Cidea mRNA response to ER stress in cell culture. Normally, Cidea gene expression is enriched in adipocytes and other cell types that contain abundant lipid droplets (9), so it initially seemed surprising to detect an increase in Cidea in the ER-stressed thyroid gland. Unfortunately, we could find no reliable antibody to examine endogenous CIDE-A protein expression in mouse tissues. Thus, to investigate the effects of ER stress on Cidea expression specifically in thyrocytes, we began by challenging the PCCL3 (rat thyrocyte) cell line with either high-dose thapsigargin (Thaps, 1 μM) for 1 hour or low-dose Thaps (0.1 μM) for 2 days. Both ER stress challenges resulted in a dramatic increase in Cidea mRNA levels ( Figure 3, A and B). Similar responses in PCCL3 cells were seen even after 6 hours of TSH withdrawal (data not shown); moreover, Thaps treatment induced the expression of BiP, Chop, and Cidea in Neuro2A cells (a neuronal cell line, Figure 4A), and similar results with Thaps (and another ER stress agent, tunicamycin; TUN) were also observed in NRK cells (a kidney cell line, Figure 4B). Thus, the effect of acute ER stress to stimulate Cidea mRNA levels is not a feature that requires ongoing TSH stimulation of thyrocytes, nor is it even unique to thyrocytes.
As the effect of ER stress on Cidea mRNA (in thyrocytes or other cell types) had not previously been noted to our knowledge, we examined the effect of low-dose actinomycin D (5 ng/mL, for partial inhibition of transcription) on the Cidea mRNA response to ER stress in PCCL3 cells. When challenged with a low dose JCI Insight 2021;6(7):e143980 https://doi.org/10.1172/jci.insight.143980 of TUN (0.1 μg/mL), we noted that BiP and Chop mRNAs that were increased at 12 hours became further increased at 24 hours, and, as expected, responses at both time points were inhibited by partial transcriptional blockade ( Figure 5, A and B). In contrast, the level of Cidea mRNA by low-dose TUN did not rise between 12 and 24 hours ( Figure 5C, red arrow) and was actually augmented by partial transcriptional blockade ( Figure  5C). We induced ER stress with cyclopiazonic acid (CPA, a reversible sarco-endoplasmic reticulum Ca 2+ -ATPase pump inhibitor) and then added high-dose actinomycin D (5 μg/mL) to block all transcription, and we observed that whereas the ER-stressed PCCL3 cells maintained Cidea mRNA levels between 3 and 12 hours, washout of CPA to reverse ER stress led to a lowering of Cidea mRNA (Supplemental Figure 3B). These data suggested that Cidea mRNA half-life was shorter in the absence of ER stress and stabilized in the presence of ER stress. If high-dose actinomycin D was added prior to the addition of TUN to induce ER stress, then the Cidea mRNA elevation was prevented (data not shown). However, when high-dose actinomycin D was added 12 hours after initiating an ER stress challenge (with low-dose TUN), induction of BiP and Chop mRNAs was profoundly suppressed, but the stress-induced elevation of Cidea mRNA was not inhibited ( Figure 5D). Together, these data suggest that the initial rapid increase of Cidea mRNA in response to acute ER stress includes both transcription and mRNA stabilization.
To examine the relationship of the classical ER stress response to Cidea mRNA, we performed individual siRNA-mediated knockdowns of each of the 3 ER stress sensors in PCCL3 cells. Notably, elevation of Cidea mRNA in response to acute ER stress (TUN, 24 hours) was not suppressed by knockdown of inositol-requiring enzyme-1 (Ire1) or protein kinase R-like ER kinase (Perk) and was actually augmented by knockdown of activating transcription factor 6 (Atf6) (Figure 6), and these effects could already be detected within 12 hours after addition of TUN to initiate ER stress (Supplemental Figure 4).

Examination of CIDE-A protein and cell death in ER-stressed thyrocytes.
A relationship between CIDE-A and cell death has been repeatedly noted since its initial discovery (8, 10) yet remains poorly understood. We used pTREtight to engineer PCCL3 (rat thyroid) cells to express mouse CIDE-A-myc upon induction by doxycycline (DOX) rather than by ER stress ( Figure 7A). In these engineered cells, addition of either TUN (for 2 days) or DOX (for 3 days) induced death in only a small fraction (<10%) of cells, but addition of both treatments induced cell death synergistically ( Figure 7B). Interestingly, although ER stress did not increase the recombinant Cideamyc mRNA level, ER stress did increase the recombinant CIDE-A-myc protein level ( Figure 7C). By immunofluorescence, the increased CIDE-A-myc protein in PCCL3 thyrocytes exhibited a punctate-like pattern seemingly compatible with that of lipid droplets that emerge from the ER ( Figure 8A) and could be distinguished from bulk ER bearing TG that was misfolded upon TUN exposure ( Figure 8B). Moreover, CIDE-A-myc protein level was rapidly increased upon addition of the proteasome inhibitor, MG132 ( Figure 7D). Thus, the data suggest that ER stress increases and stabilizes Cidea mRNA levels and increases CIDE-A protein accompanying increased cell death, which could be limited in cells with siRNA-mediated knockdown of endogenous Cidea (Supplemental Figure 5). It has been suggested that lipid droplets may be involved, directly or indirectly, in ER stress and cell death by promoting Ca 2+ loss from the ER (20). To begin to explore this possibility, PCCL3 cells were transfected to express the ER calcium-sensing fluorescence resonance energy transfer (FRET) probe, D4ER. From this, we observed that TUN-induced ER stress was accompanied by an apparent decrease in ER Ca 2+ levels, whereas Cidea knockdown appeared to protect from the ER stress-mediated loss of ER Ca 2+ (Supplemental Figure 6). It has been reported that CIDE-A may be induced as part of a noncanonical proapoptotic stress response in human pancreatic islets (21). In Akita diabetic (male) mice, pancreatic islets express misfolded mutant proinsulin (22,23), and the β cells experience both ER stress and cell death (24). As our laboratory was in possession of reverse-transcribed mRNA samples from isolated pancreatic islets of Akita mice at different stages of the disease (based on random blood glucose), we were able to examine islet Cidea mRNA levels and noted that these levels rose with worsening hyperglycemia (Supplemental Figure 7).
Limiting the rise of ER stress-mediated CIDE-A expression correlates with cell survival. We recently reported that PCCL3 thyrocytes in progressively increasing doses of TUN nevertheless adapt to ER stress and exhibit long-term survival in culture for up to a year (3). The upper 2 panels of Figure 9A are reproduced with permission from the Journal of Biological Chemistry (3); cells adapted to high-dose TUN exhibited enrichment of unglycosylated TG protein entrapped in the ER, along with an increase in the steady-state levels of KDEL-containing ER chaperones and oxidoreductases ( Figure 9B). Remarkably, however, upon exposure to the final and highest doses of TUN, the Cidea mRNA in these PCCL3 cells was suppressed below control levels and was no longer induced by high-dose TUN challenge ( Figure 9A, bottom panel).
Having adapted to high-dose TUN, the chronically ER-stressed cells were then exposed an unrelated acute ER stress challenge with Thaps (0.1 μM). Similar to that described above, at 48 hours, control PCCL3 cells induced the mRNAs of BiP, Chop, and Cidea and at 72 hours showed more than 25% cell death ( Figure  10, A-D). Remarkably, PCCL3 thyrocytes that had been adapted to chronic ER stress showed no change in their sensitivity of acute response of BiP and Chop mRNA ( Figure 10, A and B) but exhibited a dramatically blunted Cidea mRNA response ( Figure 10C) accompanied by a significantly blunted cell death response (Figure 10D). Altogether, the data suggest that in addition to its role in the biology of lipid droplets (9), CIDE-A is a novel marker that appears to be linked to a complex, noncanonical, proteotoxic ER stress response program -and limiting its stress-mediated induction has implications for cell survival.

Discussion
Protein-secreting endocrine tissues exhibit a feature through which when hormone production declines, endocrine circuitry typically activates physiological stimulation of the failing tissue, and in some instances this can lead to a compensatory expansion of endocrine cell mass. In a number of endocrine diseases, including some forms of diabetes mellitus, despite these physiological feedback loops, compensation may fail (25). In normal thyrocytes, ER stress response pathways are already significantly active in cells making physiological levels of wild-type TG (3). Patients (and animals) with congenital hypothyroidism caused by biallelic TG mutations resulting in TG protein misfolding trigger further ER stress, yielding both adaptive and cytotoxic stress responses. In such a case, cell death in the thyroid gland may limit the degree of compensatory thyroid gland growth (4). Thyrocytes are a good model for study of ER stress and stress responses because of their exceptional dedication to the synthesis of high levels of TG protein, and the maintenance of thyroid cell mass can be followed both in vivo and in vitro.
ER protein misfolding triggers activation of numerous pathways that can potentially lead to cell death from proteotoxicity (26)(27)(28), which underlies the pathogenesis of various neurodegenerative diseases and disorders of the endocrine system (29). A commonly held view is that when cells experience chronic unremitting ER stress, one or more cell-autonomous mechanisms activate cell death (30). Chronic ER stress-mediated cell death has been proposed to be triggered by a variety of alternative mechanisms, including persistently increased CHOP activity (31), ATF4 activity (32), hyperactivated IRE1 (33), JNK activation (34), ER membrane leakage (35), and others. Here, we suggest that additionally, ER stress-mediated cell death might be  linked to an inability to suppress the expression of prodeath gene products; conversely, survival might be linked to successful suppression or repression of such genes.
We know that TG cog/cog thyrocytes experience chronic unremitting ER stress (with a swollen ER accompanied by a markedly activated ER stress response), yet the thyrocytes continue to proliferate (3); however, thyroidal expression of the TG rdw transgene in TG cog/cog mice (without changing total TG mRNA or the degree of hypothyroidism, and exhibiting comparable levels of canonical ER stress markers) increases the prevailing level of thyrocyte cell death, which correlates with unsuppressed expression of Cidea mRNA in thyroid tissue ( Figure 2). The effect of acute ER stress to stimulate Cidea mRNA levels is not a feature that requires ongoing TSH stimulation of thyrocytes, nor is it a thyrocyte-specific feature. Indeed, Cidea transcripts are known to be abundant in white and brown fat as well as mammary tissue; the gene is also expressed at varying levels in heart, lung, skeletal muscle, spleen, thymus, lymphatic tissue, liver, stomach, brain, oocytes, placenta, kidney (36)(37)(38)(39), as well as colorectal and prostate cancer cells and endometrial (40) and ovarian cancer cells (41). In our experiments, to exclude any contribution of adipocytes or other cells that reside within the thyroid gland (42), we directly examined the PCCL3 thyrocyte cell line (Figure 3) (or other cell types; Figure  4) to examine the Cidea response to ER stress.
Cidea is not a classic ER stress response gene (being devoid of an apparent ER stress element or unfolded protein response element in its promoter), and thyrocyte ER stress regulation of Cidea exhibits several interesting features. Most notably, separate from the transcriptional regulation that has been reported in other cell types (43), Cidea mRNA levels in thyrocytes appear stabilized in the presence of ER stress ( Figure 5), while Atf6 expression under ER stress appears to limit overshoot of Cidea mRNA ( Figure 6). Clearly, much more needs to be learned about how Cidea regulation fits within the broader transcriptomic response to acute proteotoxic ER stress.
Beyond this, in PCCL3 thyrocytes, acute ER stress also elevates intracellular levels of the CIDE-A-myc protein, which otherwise (in unstressed cells) appears to be susceptible to rapid proteasomal degradation (Figure 7, C and D). In addition to acute ER stress enhancing CIDE-A distribution to intracellular organelles suggestive of lipid droplets that emanate from the ER (Figure 8) and influencing ER calcium homeostasis (Supplemental Figure 6), acute ER stress also stimulates cell death ( Figure 7B). Conversely, whereas Cidea mRNA appears to increase in Akita mouse pancreatic islets as the animals progress into worsening diabetes (Supplemental Figure   Figure 6. Acute ER stress response after siRNA-mediated knockdown of Ire1, Atf6, or Perk in PCCL3 cells. Replicate wells of PCCL3 were transfected with 30 nM siRNA duplexes for knockdown of Ire1, Atf6, or Perk or scrambled/noncoding oligonucleotides. At 24 hours after oligofection, the cells were treated ± TUN (0.1 μg/mL). After an additional 24 hours, all cells were lysed, and mRNA levels were measured by qRT-PCR for Ire1, Atf6, and Perk (A; *P < 0.05 versus scrambled oligo) and Cidea (B). The data show the mean ± SD (from 3 independent experiments in which each sample had 3 independent biological replicates); *P < 0.05 versus TUN-treated cells transfected with scrambled oligonucleotide (1-way ANOVA).
JCI Insight 2021;6(7):e143980 https://doi.org/10.1172/jci.insight.143980 7), Cidea knockdown has been reported to improve survival in ER-stressed pancreatic β cells intoxicated with palmitate (44). CIDE-A-mediated inhibition of cell survival may relate to its downregulation of AMP-kinase activity (45,46); indeed, increased AMP-kinase activity is thought to be linked to thyrocyte survival in the face of ER stress (3). Remarkably, in contrast with reports of mouse embryonic fibroblasts (47), PCCL3 cells adapted to chronic ER stress unexpectedly did not limit their classical BiP or Chop response when presented with a new, unrelated, acute ER stress challenge (Figure 10, A and B). Rather, Cidea mRNA was no longer elevated above baseline, and Cidea exhibited a blunted response to an unrelated acute ER stress challenge ( Figure 9A and Figure 10C) accompanied by resistance to ER stress-mediated cell death ( Figure 10D).
Importantly, noncanonical mechanisms regulating the transcriptome in response to ER stress have been increasingly recognized, including active suppression of 25%-50% of all ER stress-responsive mRNAs (48). It is highly likely that Cidea is regulated as part of a transcriptional and posttranscriptional network of activities, the dissection of which will require further detailed biological and bioinformatics analyses to help account for ER stress-related cell survival and cell death (48).
Generation of TG cog/cog mice bearing transgenic TG rdw . For thyroid-specific expression, a linearized transgene consisting of the bovine TG promoter (18) immediately upstream of the full-length mouse TG ORF encoding rdw-TG plus a triple-myc tag (i.e., rdwTg3xMyc, ref. 15) was injected into B6SJLF2 eggs, obtained by the mating of B6SJLF1 mice (The Jackson Laboratory stock number 100012). Potential transgenic founder animals were screened by PCR of tail snip DNA to generate a 284 bp fragment in the presence of the TG rdw transgene. Eleven PCR + founders were repeatedly backcrossed to C57BL/6 and TG cog/cog mice from The Jackson Laboratory; the 2 strongest expressing lines (by anti-myc Western blotting) were retained. Hypothyroid breeders were given supplemental T 4 in drinking water provided ad libitum, and supplemental T 4 was withdrawn at weaning. Animal use was performed in accordance with the University of Michigan's IACUC.
TUNEL staining. TUNEL staining (MilliporeSigma ApopTag In Situ Apoptosis Detection Kit) of mouse thyroid was performed in tissue sections, and nuclei were counterstained with DAPI. Quantitation of images was performed both manually and by using ImageJ (NIH) with the ITCN plugin.
Thyroid gland size measurement. Small animal ultrasound (VisualSonics 770) was performed on live anesthetized mice to assess thyroid lobe area (volume data were also obtained but were proportional to area). (Ultrasound-measured areas were ultimately confirmed in euthanized mice by 2-dimensional digital photography of the dissected neck with a calibrated size marker. Both sets of measurements were found to be more reproducible than wet weight of dissected thyroid tissue.) Cell death screen. A mouse apoptosis PCR array (SABiosciences PAMM-012) was used to prescreen 4 TG cog/cog thyroidal reversed-transcribed RNA samples plus 4 more bearing the TG rdw transgene, along with positive and negative control genes and plate controls to yield a small data set. Average Ct values were used to calculate the relative mRNA expression of 86 survival or death genes (analysis performed by the University of Michigan Bioinformatics Core; complete data set shown in Supplemental Figure 3A).
Cell death in culture. Twenty thousand PCCL3 cells were plated in each well (96-well plates) 24 hours before each treatment. Cells were treated as indicated, and both dead cells and total cells were measured by the CytoTox-Glo Cytotoxicity Assay Kit (Promega).
Statistics. For analyses of 3 or more groups, 1-way ANOVA with Tukey's post hoc test was performed, with data shown as mean ± SD. This included Xbp1 splicing, TUNEL staining, general qPCR (mRNA levels) and Cidea mRNA levels, cell death by CytoTox-Glo assay, and FRET ratio. Ultrasound (thyroid area) between the 2 animal groups was analyzed by 2-tailed t test. For the screening array plate, the difference between average Figure 10. PCCL3 cells that ultimately adapt to chronic TUN exhibit normal acute BiP and Chop response to Thaps but a blunted Cidea response. PCCL3 cells from Figure 9 adapted to grow continuously ± TUN (300 ng/mL) were acutely challenged with Thaps (100 nM) for 48 hours. The cells were then lysed and mRNA levels were measured by qRT-PCR for BiP (A), Chop (B), or Cidea (C). In parallel, cell death was measured at 72 hours by CytoTox-Glo (D). The data show the mean ± SD (from 3 independent experiments in which each sample had 3 independent biological replicates for qPCR and 6 for CytoTox-Glo assay); *P < 0.05 for the Thaps response in cells adapted to chronic ER stress versus cells that had not been previously adapted to chronic ER stress (1-way ANOVA).
JCI Insight 2021;6(7):e143980 https://doi.org/10.1172/jci.insight.143980 Ct values of each tested gene was compared with that of control (housekeeping) genes to yield log fold change, and the University of Michigan Bioinformatics Core statistically compared the 2 groups with P values adjusted for FDR, with a P < 0.05 considered significant.
Study approval. Animal experiments were performed under the University of Michigan IACUC-approved animal protocol PRO00008062.

Author contributions
YM and APK contributed to experimental design, data generation, data analysis, and writing of the manuscript. DL, WC, and SV contributed to data generation and validation and data analysis. LS and ML contributed to supervision and editing the manuscript. PA contributed to experimental design, data analysis, writing of the manuscript, supervision, and funding. All authors discussed results.